System

Abstract

A system is provided.SOLUTION: A system comprising: means for collecting public data on the Internet; means for pre-processing and analyzing the collected data; means for predicting the number of views of content based on the pre-processed and analyzed data; and means for providing a prediction result to a user.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The technology of the present disclosure relates to a system. [Background technology]

[0002] Patent document 1 discloses a persona chatbot control method performed by at least one processor, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to a description of the chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-180282 Summary of the Invention [Problem to be solved by the invention]

[0004] Traditionally, revenue models for content such as articles and videos relied heavily on the number of views. However, there were no objective indicators for evaluating the "substance" of content and selecting the optimal expression and structure, so they had to rely on experience and track record. This made it difficult to create new hit content, and the re-production of past successes became the norm. Furthermore, the lack of a scientific approach to maximizing advertising revenue made it difficult to create effective content. Furthermore, the practical impossibility of applying traditional A / B testing methods to content made it difficult to develop an effective content strategy. [Means for solving the problem]

[0005] The present invention provides a system for efficiently collecting public data on the Internet and preprocessing and analyzing it. Specifically, the system includes means for performing the necessary preprocessing on the collected data and extracting content features using natural language processing technology. Next, the system provides means for accurately predicting the number of content views using machine learning and predictive algorithms based on the preprocessed and analyzed data. By providing users with the prediction results in real time, users can make data-driven decisions for effective content creation. Furthermore, the system includes means for transmitting content data entered by users to a server and displaying the prediction results on a terminal, allowing users to interactively select optimal content.

[0006] "Public data on the Internet" refers to information such as web pages, news sites, and social media that is accessible to the general Internet user and does not require specific authentication or authorization.

[0007] "Means of collection" refers to software or hardware mechanisms for automatically or semi-automatically obtaining public data on the Internet.

[0008] "Preprocessing" refers to the process of removing unnecessary information from collected data and converting it into a format suitable for analysis.

[0009] "Means for analysis" refers to the techniques and algorithms used to analyze the pre-processed data and extract useful information and features.

[0010] "Predictive methods" refers to machine learning models or predictive algorithms used to estimate the number of views of content based on collected and analyzed data.

[0011] "Means for providing predicted results" refers to a system or interface for notifying users of the predicted results of the number of views in real time.

[0012] "User" refers to a person or organization that creates content such as articles or videos and uses this system to predict the number of views and effectiveness of such content.

[0013] "Natural language processing technology" refers to artificial intelligence technology for analyzing text data and understanding the meaning and structure of language.

[0014] "Means of feature extraction" refers to algorithms and technologies that identify important keywords, topics, and other features from the content of text or videos.

[0015] "Machine learning" refers to an artificial intelligence technology that uses data to learn patterns and relationships and make predictions and classifications.

[0016] A "predictive algorithm" refers to a computational method or procedure for predicting future outcomes based on given data.

[0017] "Data-driven" refers to an approach that makes decisions through data-based analysis and inference. [Brief explanation of the drawings]

[0018] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7]FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0019] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0020] First, the terms used in the following description will be explained.

[0021] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0022] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0023] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0024] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0025] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0026] [First embodiment]

[0027] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0028] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0029] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0030] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0031] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0032] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0033] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0034] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0035] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0036] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0037] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0038] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0039] Overall system overview

[0040] This invention is a system that collects, preprocesses, and analyzes public data on the Internet and predicts the number of views of content based on the collected data. The system is mainly composed of three entities: a server, a terminal, and a user.

[0041] Server Processing

[0042] The server has the following functions:

[0043] 1. Data Collection

[0044] The server uses web scraping tools and APIs to collect public data from the internet.

[0045] Collections include news sites, blogs, social media posts, and more.

[0046] 2. Data Preprocessing

[0047] The server performs text cleaning on the collected data, specifically removing HTML tags, normalizing special characters, and filtering out irrelevant data.

[0048] The cleaned data is then tokenized and converted into a format suitable for analysis.

[0049] 3. Analysis and feature extraction

[0050] The server uses natural language processing (NLP) techniques to extract features from the text data, including keywords, topics, and sentiment analysis results.

[0051] The server also performs frame analysis and speech-to-text conversion on video data to extract similar features.

[0052] 4. Training the predictive model and making predictions

[0053] The server uses this data to train machine learning models, which use algorithms such as regression analysis and neural networks.

[0054] The trained model is used to predict the number of views for different content.

[0055] 5. Providing prediction results

[0056] The server generates prediction results and provides them to users in real time, including predicted views, diffusion forecasts, and suggestions for optimal presentation and composition.

[0057] Terminal handling

[0058] The terminal provides an interface for users to access the system and input content.

[0059] 1. Content input

[0060] The user enters the title, text, tags, metadata, etc. of the article or video.

[0061] 2. Send a prediction request

[0062] The terminal converts the input content data into JSON format and sends it to the server.

[0063] The prediction results returned from the server are received and displayed to the user.

[0064] User operations

[0065] The user accesses the system through a terminal and optimizes the content in the following steps.

[0066] 1. Content Creation

[0067] The user uses the device to create new article or video content. For example, the user might enter the following as the title: "Predicting the future with AI" and the text: "It is now possible to predict the future using AI technology. In particular, technology that can predict the number of content views with high accuracy is attracting attention."

[0068] 2. Prediction Request

[0069] The user clicks the button to request a prediction of the number of views of the content, and transmits the data to the server.

[0070] The server performs the analysis and predictions and sends the results back to the device.

[0071] 3. Check and correct the prediction results

[0072] The user can check the prediction results displayed on their device, such as "Expected number of views: 1500 PV" or "Recommended title: 'How to read the future with AI technology'."

[0073] The user can then modify the title and body of the content based on the prediction results and publish the optimized content.

[0074] Specific examples

[0075] As a concrete example, we will show a series of steps in which a user creates a blog article, and then uses a server to predict and optimize the number of views of that article.

[0076] 1. A user creates a blog post using a device. The title is "How to predict the future with AI technology" and the body is "It is possible to predict the future using AI technology..."

[0077] 2. The user clicks the "View Count Forecast" button, and the device sends the content data to the server.

[0078] 3. The server analyzes the data and uses a predictive model to predict the number of views.

[0079] 4. The server generates a prediction result and sends it back to the device. For example, "Expected number of views: 1500 PV" and "Recommended title: 'How to predict the future with AI'."

[0080] 5. The user checks the prediction results on their device, changes the title to the recommended title, and publishes the article.

[0081] Through the system of the present invention, users can create effective content based on scientific evidence, which can help create new hit content.

[0082] The processing flow will be explained below.

[0083] Step 1: Data collection

[0084] The server periodically collects public data from the internet using web scraping tools and APIs.

[0085] The server retrieves text and media data from news sites, blogs, social media, etc. and stores it in storage.

[0086] Step 2: Data Preprocessing

[0087] The server removes HTML tags from the collected data, normalizes special characters, and converts them into clean text data.

[0088] The server filters out irrelevant data and noise and converts it into a format suitable for analysis.

[0089] Step 3: Feature extraction

[0090] The server uses natural language processing (NLP) technology to extract features such as keywords, topics, and sentiment analysis results from the text data.

[0091] The server also performs frame analysis and speech-to-text conversion on video data to extract similar features.

[0092] Step 4: Training the predictive model

[0093] The server uses the pre-processed and feature-extracted data to create a training dataset for machine learning.

[0094] The server trains a view prediction model using a regression analysis model or a neural network.

[0095] The server stores the trained model and makes it available for real-time predictions.

[0096] Step 5: Enter content

[0097] The user uses the device interface to input the title, text, tags, metadata, etc. of the article or video.

[0098] Step 6: Send a prediction request

[0099] When the user clicks the "View Count Prediction" button, the device converts the content data into JSON format and sends it to the server.

[0100] Step 7: Data analysis and prediction

[0101] The server analyzes the received content data and performs pre-processing and feature extraction.

[0102] The server uses a trained prediction model to predict the number of views of the content.

[0103] Step 8: Providing prediction results

[0104] The server generates prediction results and returns them to the user, including predicted views, diffusion forecasts, and suggestions for optimal presentation and composition.

[0105] Step 9: Check and correct the predictions

[0106] The user checks the prediction results received on their device, such as "Expected number of views: 1500 PV" and "Recommended title: 'How to read the future with AI technology.'"

[0107] The user can then modify the title and body of the content based on the prediction results and publish the optimized content.

[0108] Example 1

[0109] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0110] In conventional content view prediction systems, the processes from data collection to preprocessing, analysis, prediction, and result provision are often not performed consistently and efficiently. Also, bias toward specific analysis methods or models can make accurate predictions difficult. Another issue is the lack of a means to efficiently transmit content data entered by users to a server in real time and quickly return prediction results. There is a need for a more accurate and efficient content view prediction system that can resolve these issues.

[0111] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0112] In this invention, the server includes means for collecting public data on the Internet, means for preprocessing the collected data, including removing HTML tags, normalizing special characters, and filtering irrelevant data, means for tokenizing the preprocessed data and converting it into a format suitable for analysis, means for extracting features such as keywords, topics, and sentiment analysis results from the text data using natural language processing techniques, means for training a machine learning model based on the analyzed data and predicting the number of views of content, and means for providing the predicted results to users in real time. This makes it possible to efficiently collect data, perform preprocessing and analysis, accurately predict the number of views using the machine learning model, and quickly provide the results to users.

[0113] A "server" is a computer system that collects, analyzes, and predicts public data on a network and provides the results to users.

[0114] "Public data on the internet" is information that is freely accessible online, such as news sites, blogs, and social media posts.

[0115] "Preprocessing" refers to data cleaning tasks such as removing HTML tags from collected data, normalizing special characters, and filtering out irrelevant data.

[0116] "Tokenization" is the process of breaking down text data into units of words and phrases and converting them into a format suitable for analysis.

[0117] "Natural language processing technology" is a technology for extracting features such as keywords, topics, and emotions from text data.

[0118] A "machine learning model" is an algorithm or program that is trained on data to perform a specific task (in this case, predicting content views).

[0119] "Real-time" refers to a form of immediate response where the time between when a user sends data and when they receive the results is very short.

[0120] A "terminal" is a computer or device that allows a user to access the system, enter content, and view results.

[0121] "Content" is a collection of data including article or video titles, text, tags, metadata, etc.

[0122] MODE FOR CARRYING OUT THE INVENTION

[0123] This system collects public data on the Internet, preprocesses and analyzes it, and then predicts the number of views of content based on the collected data. The system is mainly composed of three entities: a server, a terminal, and a user.

[0124] Server Processing

[0125] The server performs data processing and data calculation using the following hardware and software.

[0126] 1. Data Collection

[0127] The server uses web scraping tools (e.g., BeautifulSoup, Scrapy) and APIs (e.g., Twitter API, YouTube Data API) to collect data from news sites, blogs, social media posts, etc.

[0128] Specifically, the server sends an HTTP request to the specified URL and retrieves HTML data or JSON format data.

[0129] 2. Data Preprocessing

[0130] The server performs text cleaning such as removing HTML tags, normalizing special characters, and filtering out irrelevant data. For example, it uses BeautifulSoup to remove HTML tags and regular expressions to normalize special characters.

[0131] The cleaned data is then tokenized and converted into a format suitable for analysis, for example by splitting the text into words using Python's NLTK library.

[0132] 3. Analysis and feature extraction

[0133] The server analyzes the text data using natural language processing (NLP) techniques to extract features such as keywords, topics, and sentiment analysis results. For example, it uses the spaCy library to extract keywords and topics, and calculates sentiment scores using VADER Sentiment Analysis.

[0134] For video data, frame analysis and speech-to-text conversion are performed, for example, by using OpenCV to extract frame-by-frame features and Google Cloud Speech-to-Text to convert the speech to text.

[0135] 4. Training the predictive model and making predictions

[0136] The server trains a machine learning model based on the analyzed feature data, for example, a linear regression model from scikit-learn or a neural network from TensorFlow.

[0137] Use the trained model to predict the number of views for new content.

[0138] 5. Providing prediction results

[0139] The server generates prediction results in JSON format and sends them back to the user's device.

[0140] Terminal handling

[0141] The terminal provides an interface for users to access the system, input content, and check prediction results.

[0142] 1. Content input

[0143] The user uses the device to enter the title, text, tags, and metadata for the article or video.

[0144] 2. Send a prediction request

[0145] The terminal converts the input content data into JSON format and sends it to the server.

[0146] The prediction results returned from the server are received and displayed to the user.

[0147] User operations

[0148] The user accesses the system through a terminal and optimizes the content in the following steps.

[0149] 1. Content Creation

[0150] The user uses the device to create new article or video content. For example, the user might enter the following as the title: "Predicting the future with AI" and the text: "It is now possible to predict the future using AI technology. In particular, technology that can predict the number of content views with high accuracy is attracting attention."

[0151] 2. Prediction Request

[0152] The user clicks the button to request a prediction of the number of views of the content, and the terminal transmits the data to the server.

[0153] 3. Check and correct the prediction results

[0154] The server performs the analysis and predictions and sends the results back to the device. For example, results such as "Expected number of views: 1500 PV" and "Recommended title: 'How to predict the future with AI technology'" are displayed.

[0155] Users can then modify and optimize the title and body of their content based on the prediction results, and then publish the optimized content.

[0156] Through this system, users will be able to create effective content based on scientific evidence, helping to create new hit content.

[0157] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0158] Step 1: Data collection

[0159] Input: URLs of public data on the internet (news sites, blogs, social media posts, etc.)

[0160] How it works: The server sends an HTTP request to a URL to retrieve data in HTML or JSON format, using a web scraping tool (e.g., BeautifulSoup, Scrapy) or an API (e.g., Twitter API, YouTube Data API).

[0161] Output: Collected HTML data and JSON format data

[0162] Step 2: Data Preprocessing - Removing HTML Tags and Normalizing Special Characters

[0163] Input: Collected HTML data or JSON format data

[0164] What it does: The server uses BeautifulSoup to strip HTML tags and regular expressions to normalize special characters.

[0165] Output: Cleaned text data

[0166] Step 3: Data Preprocessing - Data Filtering

[0167] Input: Cleaned text data

[0168] What it does: The server removes data that does not contain certain keywords, for example, filtering out unwanted advertisements or spam posts.

[0169] Output: filtered text data

[0170] Step 4: Tokenization

[0171] Input: filtered text data

[0172] Specific operation: The server uses Python's NLTK library to split the text data into words.

[0173] Output: Tokenized text data

[0174] Step 5: Feature Extraction

[0175] Input: Tokenized text data

[0176] Specific operation: The server extracts features such as keywords, topics, and sentiment scores using natural language processing (NLP) techniques (e.g., spaCy, VADER Sentiment Analysis).

[0177] Output: Feature data such as keywords, topics, and sentiment scores

[0178] Step 6: Analyze the video data

[0179] Input: Video data

[0180] Specific operation: The server uses OpenCV to extract features for each frame and converts the audio into text using Google Cloud Speech-to-Text.

[0181] Output: Frame-by-frame feature data and speech-to-text data

[0182] Step 7: Train the machine learning model

[0183] Input: Feature data (keywords, topics, sentiment scores, etc.) and filtered data

[0184] Specific operation: The server trains a machine learning model based on the feature data using a linear regression model from scikit-learn or a neural network from TensorFlow.

[0185] Output: A trained machine learning model

[0186] Step 8: Predict content views

[0187] Input: New content feature data (keywords, topics, sentiment scores, etc.)

[0188] What it does: The server uses a trained machine learning model to predict the number of views for new content.

[0189] Output: Predicted number of views

[0190] Step 9: Providing prediction results

[0191] Input: Estimated number of views

[0192] Specific operation: The server formats the prediction results in JSON format and sends them to the device.

[0193] Output: Formatted prediction results (JSON format)

[0194] Step 10: Enter content

[0195] Input: Users enter content such as article or video titles, text, tags, and metadata.

[0196] What happens: The device takes these inputs and prepares the data for the prediction request.

[0197] Output: Input content data (JSON format)

[0198] Step 11: Send a prediction request

[0199] Input: Input content data (JSON format)

[0200] Specific operation: The terminal sends content data formatted in JSON format to the server via an HTTP POST request.

[0201] Output: Request sent

[0202] Step 12: Receive and display results

[0203] Input: Prediction results returned from the server (JSON format)

[0204] Specific behavior: The device receives the prediction results and displays them to the user. The device's UI components are used to visually display the results, such as the predicted number of views and recommended titles.

[0205] Output: Prediction results displayed to the user

[0206] This allows users to efficiently collect data, preprocess and analyze it, use machine learning models to accurately predict the number of views, and quickly provide the results to users.

[0207] (Application example 1)

[0208] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0209] While there are many systems that use publicly available data on the Internet to predict content view counts, most of them only provide prediction results and do not include specific optimization suggestions. As a result, users lack the information to effectively optimize titles and thumbnails based on the prediction results, and must rely on trial and error. This makes it difficult to maximize content views.

[0210] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0211] In this invention, the server includes means for collecting public data on the Internet, means for preprocessing and analyzing the collected data, means for predicting the number of views of content based on the preprocessed and analyzed data, means for optimizing the title and thumbnail of the content based on the predicted number of views, and means for providing the prediction results and optimization suggestions to users, thereby enabling users to efficiently and effectively optimize their content and maximize the number of views.

[0212] "Public data on the internet" refers to digital information that is freely accessible from networked sources, such as news sites, blogs, public API data, and social media posts.

[0213] "Methods of collection" refers to the techniques and methods used to obtain data from designated sources, including web scraping and the use of APIs.

[0214] "Preprocessing" refers to a series of operations performed to transform collected data into a format that is easier to analyze, such as text cleaning, normalizing special characters, and filtering out irrelevant data.

[0215] "Analytical means" refers to techniques and methods for analyzing collected and pre-processed data, including the application of natural language processing techniques and machine learning algorithms.

[0216] "Methods for predicting views" refers to techniques or methods for predicting the number of views of content based on analyzed data, including models such as regression analysis and neural networks.

[0217] "Optimization means" refers to technologies and methods for effectively changing or suggesting content titles and thumbnails based on the prediction results.

[0218] "Means of providing" refers to the interface and functions for notifying and displaying prediction results and optimization suggestions to users, including the user interface and notification functions.

[0219] This system effectively maximizes the number of content views by collecting, preprocessing, and analyzing public data on the Internet and providing users with prediction results and optimization suggestions. This system is mainly composed of three entities: a server, a terminal, and a user.

[0220] Server Processing

[0221] The server has the following functions:

[0222] 1. Data Collection:

[0223] The server uses web scraping tools and APIs to collect public data from the internet, such as news sites, blogs, and social media posts.

[0224] 2. Data Preprocessing:

[0225] The server performs text cleaning on the collected data, specifically removing HTML tags, normalizing special characters, and filtering out irrelevant data.

[0226] 3. Analysis and feature extraction:

[0227] The server uses natural language processing (NLP) techniques to extract features from the text data, including keywords, topics, and sentiment analysis results. It also uses the Python library Scikit-learn to vectorize the text data using TF-IDF.

[0228] 4. Views forecast:

[0229] The server trains a machine learning model based on the pre-processed and feature-extracted data to predict the number of views. It uses algorithms such as regression analysis and neural networks to build the predictive model.

[0230] 5. Optimization:

[0231] Based on the prediction results, the server generates suggestions to optimize the title and thumbnail of the content.

[0232] 6. Providing prediction results:

[0233] The server provides the prediction results and optimization suggestions to the user in real time.

[0234] Terminal handling

[0235] The terminal provides an interface for users to access the system and input content.

[0236] 1. Content Input:

[0237] Users use the device to input the title, text, tags, metadata, etc. of articles and videos.

[0238] 2. Send a prediction request:

[0239] The terminal converts the input content data into JSON format and sends it to the server.

[0240] 3. Receiving and displaying prediction results:

[0241] The terminal receives the prediction results and optimization suggestions returned from the server and displays them to the user.

[0242] User operations

[0243] The user accesses the system through a terminal and optimizes the content in the following steps.

[0244] 1. Content Creation:

[0245] Users use their devices to create new article or video content, for example by typing: "How to read the future with AI technology."

[0246] 2. Prediction request:

[0247] The user clicks the button requesting a prediction of the number of views of the content, and transmits data from the terminal to the server.

[0248] 3. Check and correct the predictions:

[0249] Users can check the prediction results and optimization suggestions displayed on their devices. For example, they may see "Expected page views: 1,500 PV" or "Recommended title: 'How to predict the future with AI technology.'" Based on this, users can revise the title and thumbnail and publish the optimized content.

[0250] Example prompt sentence:

[0251] The system collects and preprocesses data in the following way, then analyzes it to predict the number of views of the content. The specific steps are as follows:

[0252] 1. Collect news sites, blogs, and social media posts using web scraping tools and APIs.

[0253] 2. Clean the collected data and convert it into a suitable format for analysis.

[0254] 3. Extract features using natural language processing techniques.

[0255] 4. Train a machine learning model to predict viewership.

[0256] 5. Provide users with real-time prediction results.

[0257] Users can generate optimal titles and descriptions for various content to maximize views. For example, a video title like "How to Read the Future with AI Technology" is expected to attract many viewers according to the predictive model.

[0258] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0259] Step 1: Data collection

[0260] The server collects public data from news sites, blogs, social media posts, and other sources on the internet. To do this, it uses a web scraping tool or API. Specifically, it sends HTTP requests to specified URLs and parses the resulting HTML data to extract usable information. A list of URLs is given as input, and text data is obtained as output.

[0261] Step 2: Data Preprocessing

[0262] The server preprocesses the collected text data by removing HTML tags, normalizing special characters, and filtering irrelevant data. Raw text data is given as input, and clean text data is obtained as output. Regular expressions and text cleaning libraries are used in this process.

[0263] Step 3: Feature extraction

[0264] The server extracts features using the preprocessed text data. To do this, it uses natural language processing (NLP) techniques. Specifically, it uses the Python library Scikit-learn to perform TF-IDF vectorization. Clean text data is given as input, and vectorized feature data is obtained as output.

[0265] Step 4: Train the machine learning model

[0266] The server uses the feature data to train a machine learning model. This uses algorithms such as regression analysis and neural networks. Specifically, it uses the LinearRegression model from Scikit-learn. Vectorized feature data and label data (past view counts) are given as input, and the trained model is obtained as output.

[0267] Step 5: Predicting views

[0268] The server uses the trained machine learning model to predict the number of views for new content data provided by the user. The server receives feature data of the new content as input and obtains the predicted number of views as output.

[0269] Step 6: Optimization suggestions

[0270] The server generates optimal title and thumbnail suggestions based on the prediction results, taking into account past data on similar methods that have generated large numbers of views. The predicted number of views and content feature data are given as input, and the output is an optimized suggestion.

[0271] Step 7: Providing prediction results and optimization suggestions

[0272] The server provides the prediction results and optimization suggestions to the user in real time. This information is displayed to the user through the terminal. The prediction results and optimization suggestions are given as input and displayed to the user as output.

[0273] Step 8: Optimize your content for users

[0274] The user modifies the title and thumbnail of the content based on the provided prediction results and optimization suggestions, which effectively maximizes the number of views of the content. The prediction results and optimization suggestions are given as input, and the optimized content is obtained as output.

[0275] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0276] Overall system overview

[0277] This system collects, preprocesses, and analyzes public data on the Internet, and then predicts the number of views of content based on that data. Furthermore, it incorporates an emotion engine that recognizes user emotions, allowing for predictions that take emotion data into account. This system is primarily comprised of three components: a server, a terminal, and a user.

[0278] Server Processing

[0279] The server has the following functions:

[0280] 1. Data Collection

[0281] The server periodically collects public data from the internet using web scraping tools and APIs.

[0282] The server retrieves text and media data from news sites, blogs, social media, etc. and stores it in storage.

[0283] 2. Data Preprocessing

[0284] The server removes HTML tags from the collected data, normalizes special characters, and converts them into clean text data.

[0285] The server filters out irrelevant data and noise and converts it into a format suitable for analysis.

[0286] 3. Feature Extraction

[0287] The server uses natural language processing (NLP) technology to extract features such as keywords, topics, and sentiment analysis results from the text data.

[0288] The server also performs frame analysis and speech-to-text conversion on video data to extract similar features.

[0289] 4. Analysis by Emotion Engine

[0290] The server uses an emotion engine to recognize emotions from text data about users and content, for example classifying positive, negative, and neutral emotions.

[0291] The server incorporates the emotion data into the pre-processing and feature extraction data.

[0292] 5. Training the predictive model and making predictions

[0293] The server uses the pre-processed and feature-extracted data and the sentiment data to create a training dataset for machine learning.

[0294] The server trains a view prediction model using a regression analysis model or a neural network.

[0295] The server stores the trained model and makes it available for real-time predictions.

[0296] 6. Providing prediction results

[0297] The server generates prediction results and provides them to users in real time, including predicted views, virality predictions, suggestions for optimal presentation and composition, and sentiment-based recommendations.

[0298] Terminal handling

[0299] The terminal provides an interface for users to access the system and input content.

[0300] 1. Content input

[0301] The user enters the title, text, tags, metadata, etc. of the article or video.

[0302] 2. Send a prediction request

[0303] The terminal converts the input content data into JSON format and sends it to the server.

[0304] 3. Check the prediction results

[0305] The terminal receives the prediction results returned from the server and displays them to the user.

[0306] User operations

[0307] The user accesses the system through a terminal and optimizes the content in the following steps.

[0308] 1. Content Creation

[0309] The user uses the device to create new article or video content. For example, the user might enter the following as the title: "Predicting the future with AI" and the text: "It is now possible to predict the future using AI technology. In particular, technology that can predict the number of content views with high accuracy is attracting attention."

[0310] 2. Prediction Request

[0311] The user clicks the button to request a prediction of the number of views of the content, and transmits the data to the server.

[0312] The server performs the analysis and predictions and sends the results back to the device.

[0313] 3. Check and correct the prediction results

[0314] The user checks the prediction results displayed on their device, such as "Expected number of views: 1500 PV," "Recommended title: 'How to read the future with AI technology,'" and "Emotion score: High."

[0315] The user can then modify the title and body of the content based on the prediction results and publish the optimized content.

[0316] Specific examples

[0317] As a concrete example, we show a series of steps in which a user creates a blog post, a server is used to predict the number of views of the post, and optimization is performed taking into account emotional data.

[0318] 1. A user creates a blog post using a device. The title is "How to predict the future with AI technology" and the body is "It is possible to predict the future using AI technology..."

[0319] 2. The user clicks the "View Count Forecast" button, and the device sends the content data to the server.

[0320] 3. The server analyzes the data and extracts emotion data using an emotion engine.

[0321] 4. The server uses the trained prediction model to predict the number of views and generates a prediction result.

[0322] 5. The server sends the prediction results back to the device. For example, "Expected number of views: 1500 PV", "Recommended title: 'How to predict the future with AI'", "Sentiment score: High".

[0323] 6. The user checks the prediction results on their device, changes the title to the recommended title, and publishes the optimized article.

[0324] Through the system of the present invention, users can create effective content based on scientific evidence that takes into account emotional data, thereby supporting the creation of new hit content.

[0325] The processing flow will be explained below.

[0326] Step 1: Data collection

[0327] The server periodically collects public data from the internet using web scraping tools and APIs.

[0328] The server retrieves text and media data from news sites, blogs, social media, etc. and stores it in storage.

[0329] Step 2: Data Preprocessing

[0330] The server removes HTML tags from the collected data, normalizes special characters, and converts them into clean text data.

[0331] The server filters out irrelevant data and noise and converts it into a format suitable for analysis.

[0332] Step 3: Feature extraction

[0333] The server uses natural language processing (NLP) technology to extract features such as keywords, topics, and sentiment analysis results from the text data.

[0334] The server also performs frame analysis and speech-to-text conversion on video data to extract similar features.

[0335] Step 4: Analysis by Emotion Engine

[0336] The server uses an emotion engine to recognize emotions from text data about users and content.

[0337] The server classifies and scores positive, negative, and neutral sentiment within the text data.

[0338] Step 5: Train the predictive model

[0339] The server uses the pre-processed and feature-extracted data and the sentiment data to create a training dataset for machine learning.

[0340] The server trains a view prediction model using a regression analysis model or a neural network.

[0341] The server stores the trained model and makes it available for real-time predictions.

[0342] Step 6: Enter content

[0343] The user uses the device interface to input the title, text, tags, metadata, etc. of the article or video.

[0344] Step 7: Send a prediction request

[0345] When the user clicks the "View Count Prediction" button, the device converts the content data into JSON format and sends it to the server.

[0346] Step 8: Data analysis and prediction

[0347] The server analyzes the received content data and performs pre-processing and feature extraction.

[0348] The server analyzes the emotion of the content using an emotion engine and calculates an emotion score.

[0349] The server uses a trained prediction model to predict the number of views of the content.

[0350] Step 9: Providing prediction results

[0351] The server generates and returns prediction results to the user, including expected views, virality predictions, suggestions for optimal wording and composition, and sentiment scores.

[0352] Step 10: Check and correct the prediction results

[0353] The user checks the prediction results received on their device, which may include, for example, "Expected number of views: 1500 PV," "Recommended title: 'How to read the future with AI technology,'" and "Emotion score: High."

[0354] The user can then modify the title and body of the content based on the prediction results.

[0355] The user publishes the optimized content.

[0356] Example 2

[0357] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0358] Conventional systems that use public data on the Internet to predict the number of views of content are limited to preprocessing and analysis of collected data, and it is difficult to make predictions that take user emotions into account. As a result, prediction accuracy is low and it is difficult to create effective content.

[0359] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for collecting public data on the Internet, means for preprocessing and analyzing the collected data, means for predicting the number of views of content based on the preprocessed and analyzed data, means for extracting and analyzing emotion data using an emotion engine that recognizes the user's emotions, and means for providing the prediction result to the user. This makes it possible to predict the number of views of content with high accuracy, taking the user's emotions into consideration.

[0360] "Public data on the Internet" refers to all information, including text data and media data, that is publicly available on the Internet.

[0361] "Means of collection" refers to the function of regularly obtaining publicly available data on the Internet using web scraping tools and APIs.

[0362] "Preprocessing" refers to the process of removing HTML tags and normalizing special characters from collected data to convert it into clean text data.

[0363] "Means of analysis" refers to the function of analyzing preprocessed data using natural language processing technology, etc., and extracting necessary features and information.

[0364] "Means for predicting the number of content views" refers to a function that uses a machine learning model to predict the number of content views based on pre-processed and analyzed data.

[0365] An "emotion engine" is software or a service for analyzing user emotions, and has the ability to recognize positive, negative, and neutral emotions from text data.

[0366] "Emotion data" is the result data of emotions related to users and content extracted by the emotion engine.

[0367] The "means for providing predicted results" refers to the function of the server to generate predicted results and provide them to the user in real time.

[0368] "Natural language processing technology" is a technology for analyzing text data and extracting features, and includes keyword extraction, topic modeling, sentiment analysis, etc.

[0369] "Content data entered by the user" refers to information such as the title, text, tags, and metadata of articles and videos that the user sends to the system via their terminal.

[0370] "Server" refers to a computer system that centrally handles functions such as data collection, preprocessing, analysis, emotion recognition, and view count prediction.

[0371] "Terminal" refers to the device or application that a User uses to access the System and input Content.

[0372] This system collects, preprocesses, and analyzes public data on the Internet, and predicts the number of views of content based on that data. Furthermore, it combines an emotion engine that recognizes user emotions, allowing for predictions that take emotion data into account.

[0373] Server functions and processing contents

[0374] 1. Data Collection

[0375] The server uses web scraping tools (e.g., BeautifulSoup, Scrapy) and APIs to collect public data from news sites, blogs, social media, etc. For example, article text from news sites and social media posts are stored in a database. The server periodically collects this data and stores it in storage.

[0376] 2. Data Preprocessing

[0377] The server converts the collected data into clean text data by removing HTML tags and normalizing special characters. Specifically, it uses BeautifulSoup to remove unnecessary tags and format the data in a unified format. It also filters out advertisements and irrelevant parts to create a dataset suitable for analysis.

[0378] 3. Feature Extraction

[0379] The server uses natural language processing (NLP) techniques (e.g., spaCy, NLTK) to extract features such as keywords, topics, and sentiment analysis results from text data. For example, it performs topic modeling and sentiment scoring. For video data, it performs frame analysis using OpenCV and speech-to-text conversion using Google Speech-to-Text to extract similar features.

[0380] 4. Analysis by Emotion Engine

[0381] The server analyzes the user's emotions using an emotion engine (e.g., Natural Language Understanding), which classifies positive, negative, and neutral emotions from the text data and integrates the emotion data into a feature dataset.

[0382] 5. Training the predictive model and making predictions

[0383] The server uses the preprocessed and feature-extracted dataset to train a machine learning model (e.g., TensorFlow, PyTorch). The trained predictive model is used to predict the number of views of new content in real time.

[0384] 6. Providing prediction results

[0385] The server uses the trained model to generate predictions and provide them to the user, such as generating JSON data containing predicted views and sentiment-based recommendations, and sending it back to the device.

[0386] User and Device Capabilities

[0387] 1. Content input

[0388] The user uses the device to enter the title, text, tags, and metadata of a new article or video. For example, the title might be "Predicting the future with AI" and the text might be "AI technology makes it possible to predict the future...".

[0389] 2. Send a prediction request

[0390] When the user clicks the "View Count Prediction" button, the device converts the entered content data into JSON format and sends it to the server.

[0391] 3. Check the prediction results

[0392] The device receives the prediction results sent from the server and displays them to the user, such as "Expected number of views: 1500 PV," a recommended title "How to predict the future with AI technology," and an emotion score of "High."

[0393] 4. Modify and optimize your content

[0394] The user checks the prediction results and modifies the content as necessary, for example, changing the title to a recommended title or adjusting the wording of the body text. The optimized content is then published by the user.

[0395] Examples and prompts

[0396] As a concrete example, we show a series of steps in which a user creates a blog post, a server is used to predict the number of views of the post, and optimization is performed taking into account emotional data.

[0397] Title: "How to predict the future with AI technology"

[0398] Body text: "AI technology is making it possible to predict the future. In particular, technology that can predict the number of content views with high accuracy is attracting attention."

[0399] An example prompt might be, "What is the expected number of views for this article?"

[0400] Through this system, users will be able to create effective content based on scientific evidence, helping to create new hit content.

[0401] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0402] Step 1: Data collection

[0403] Input: The server retrieves public data from the internet (e.g., news articles, blog posts, social media posts, etc.) from a specified URL or API.

[0404] What it does: The server collects public data using web scraping tools (e.g., BeautifulSoup, Scrapy) or APIs. For example, it scrapes articles from news sites with BeautifulSoup and retrieves social media posts through APIs.

[0405] Output: The collected raw data is stored in the server's storage. It includes files in various formats, such as text data, image data, and audio data.

[0406] Step 2: Data Preprocessing

[0407] Input: Raw data collected by the server.

[0408] What it does: The server uses BeautifulSoup to strip HTML tags, normalize special characters, and perform noise filtering to remove irrelevant data, such as removing ads and script tags from HTML and resolving encoding issues.

[0409] Output: Clean text data and structured data. For example, ads and irrelevant information are removed and a normalized text file is output.

[0410] Step 3: Feature extraction

[0411] Input: Clean text data.

[0412] Specific operation: The server uses natural language processing (NLP) techniques (e.g., spaCy, NLTK) to extract features such as keywords, topics, and sentiment analysis results. For video data, it performs frame analysis using OpenCV and converts the audio to text using Google Speech-to-Text.

[0413] Output: A dataset of extracted features, including a keyword list and sentiment scores from text data, and important scenes extracted from video frames.

[0414] Step 4: Analysis by Emotion Engine

[0415] Input: Extracted feature dataset.

[0416] Specific operation: The server uses an emotion engine (e.g., Natural Language Understanding) to classify the emotion of each piece of text data and calculate an emotion score, for example, classifying it as "positive," "negative," or "neutral."

[0417] Output: A feature dataset with sentiment data added, e.g., sentiment scores for each text data.

[0418] Step 5: Train the predictive model and make predictions

[0419] Input: Preprocessed and feature extracted dataset.

[0420] What it does: The server uses machine learning models (e.g., TensorFlow, PyTorch) to train regression analysis or neural network models using this data, and then uses the trained models to predict the number of views for new content.

[0421] Output: The trained prediction model and prediction results, such as predicted views and suggested titles for new articles.

[0422] Step 6: Providing prediction results

[0423] Input: JSON data containing prediction results.

[0424] Specific operation: The server generates a prediction result and sends it back to the device. For example, "Expected number of views: 1500 PV" or "Recommended title: 'How to predict the future with AI technology'".

[0425] Output: The prediction result is sent to the device so that the user can check it.

[0426] Step 7: User input and prediction request

[0427] Input: User-generated content data (article title, body, tags, metadata, etc.).

[0428] Specific operation: The user inputs content through the terminal and clicks the "View Count Prediction" button. The terminal converts the content data into JSON format and sends it to the server.

[0429] Output: The content data sent to the server.

[0430] Step 8: Check the prediction results and modify the content

[0431] Input: Prediction results sent from the server.

[0432] Specific operation: The user checks the prediction results on the device and modifies the content as necessary. For example, the title can be changed to "How to read the future with AI technology" based on the prediction results.

[0433] Output: The optimized content is published.

[0434] As described above, data input, data processing, data calculation, and output are performed in each processing step, enabling users to create effective content by utilizing highly accurate prediction results and emotion data.

[0435] (Application example 2)

[0436] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0437] Conventional content view count prediction systems simply predicted the number of views without considering user emotions, making it difficult to optimize content more effectively. In particular, in the advertising field, where emotion-based optimization is required, there was a lack of functionality to suggest optimal titles and content.

[0438] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0439] In this invention, the server includes means for collecting public data on the Internet, means for preprocessing and analyzing the collected data, means for predicting the number of views of content based on the preprocessed and analyzed data, means for recognizing user emotions using a sentiment analysis engine and making predictions taking the emotional data into consideration, means for providing the prediction results and optimization suggestions to the user, and means for generating optimal titles and content revision suggestions for advertising content, thereby enabling the prediction and optimization of the number of views of advertising content that reflects user emotions.

[0440] "Public data on the internet" refers to information that is freely accessible on websites, blogs, social media platforms, etc.

[0441] "Means of collection" refers to the process of using web scraping tools or APIs to obtain public data on a server.

[0442] "Preprocessing" refers to the process of removing unnecessary HTML tags from collected data, normalizing special characters, and converting the data into a format suitable for analysis.

[0443] "Means of analysis" refers to the process of analyzing data using natural language processing techniques and extracting features such as keywords, topics, and sentiment.

[0444] "Method for predicting content views" refers to the process of using a machine learning model to predict content views based on pre-processed and analyzed data.

[0445] "Sentiment analysis engine" refers to an algorithm that recognizes positive, negative, and neutral emotions from text data entered by the user.

[0446] "Means for making predictions taking into account emotional data" refers to a process of incorporating emotional data into a view count prediction model to make view count predictions that take into account the influence of emotions.

[0447] "Means for providing prediction results and optimization suggestions to users" refers to a process for providing users with prediction results as well as suggestions for more effective content expression and configuration.

[0448] "Means for generating optimal titles and suggested content modifications for advertising content" refers to the process of generating effective titles and suggested text modifications for advertising content based on prediction results and emotional data.

[0449] The present invention relates to a system that collects, preprocesses, and analyzes public data on the Internet and predicts the number of views of content based on that data. This system, combined with a sentiment analysis engine, performs advanced predictions that take emotional data into account and provides optimal titles and content revisions, particularly in the advertising field. A specific embodiment of this system is described below.

[0450] Server Processing

[0451] The server has the following main functions:

[0452] 1. Data Collection

[0453] The server periodically collects public data from the internet using web scraping tools and APIs.

[0454] For example, text and media data is acquired from news sites, blogs, and social media and stored in a database.

[0455] 2. Data Preprocessing

[0456] The server removes HTML tags from the collected data, normalizes special characters, and converts them into clean text data.

[0457] It also filters out irrelevant data and noise and converts it into a format suitable for analysis.

[0458] 3. Feature Extraction

[0459] The server uses natural language processing technology (e.g., spaCy or NLTK) to extract features such as keywords, topics, and sentiment analysis results from the text data.

[0460] In the case of video data, frame analysis and speech-to-text conversion are performed to extract similar features.

[0461] 4. Analysis using a sentiment analysis engine

[0462] The server uses a sentiment analysis engine (such as VADER) to classify the user's sentiment from the text data.

[0463] The classified sentiment data (positive, negative, neutral) is incorporated into the preprocessing and feature extraction data.

[0464] 5. Training the predictive model and making predictions

[0465] The server uses the preprocessed and feature-extracted data and sentiment data to train a machine learning model (e.g., Scikit-learn or TensorFlow).

[0466] The trained model is used to predict the number of views of the content.

[0467] 6. Providing prediction results and optimization suggestions

[0468] The server generates and provides the user with optimal representation and composition suggestions along with the prediction results.

[0469] Specifically, it generates optimal titles and suggested content revisions for advertising content, reflecting user emotional data.

[0470] Terminal handling

[0471] The terminal provides an interface for users to access the system and input content.

[0472] 1. Content input

[0473] Users enter the title, text, tags, metadata, etc. of the article or video.

[0474] 2. Send a prediction request

[0475] The terminal converts the input content data into JSON format and sends it to the server.

[0476] 3. Check the prediction results and optimization suggestions

[0477] The terminal receives the prediction results and optimization suggestions returned from the server and displays them to the user.

[0478] User operations

[0479] A user accesses the system through a terminal and optimizes advertising content by following the steps below.

[0480] 1. Content Creation

[0481] The user uses the device to create new advertising content, for example, by typing "We've launched a new smartwatch. Learn more!"

[0482] 2. Prediction Request

[0483] The user clicks the button to request a prediction of the number of views of the content, and transmits the data to the server.

[0484] 3. Check and modify the prediction results and optimization suggestions

[0485] Users can view the prediction results and optimization suggestions displayed on their device, such as "Expected page views: 3,000 PV," "Recommended title: 'Limited sale! Get it now,'" and "Sentiment score: High."

[0486] Users optimize and publish content based on prediction results and suggestions.

[0487] Specific examples

[0488] When a user submits an ad to the system saying, "We've released a new smartwatch. Click here for details!", the system performs sentiment analysis and predicts the number of views. The predicted results are displayed as "Sentiment score: Medium," "Expected number of views: 2000 PV," and "Recommended title: 'Limited time only! Check out the details of the limited edition smartwatch.'" The user can then change the displayed title to optimize the ad.

[0489] Prompt Sentence Examples

[0490] Prompt statement:

[0491] The ad data entered by the user is listed below. Please perform sentiment analysis and respond with a view prediction and optimal ad modification suggestions.

[0492] Ad title: New smartwatch released

[0493] Ad text: We've launched a new smartwatch. Learn more here!

[0494] Tags: smartwatch sale

[0495] Expected output:

[0496] Estimated number of views: XXXX PV

[0497] Recommended title: 'Now only! Check out the details of the limited edition smartwatch'

[0498] Sentiment score: High / Medium / Low

[0499] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0500] Step 1:

[0501] Data collection (server)

[0502] Input: Public data from the internet (e.g., news articles, blog posts, social media posts).

[0503] What it does: The server periodically collects this public data using web scraping tools and APIs and stores it in a database.

[0504] Output: The raw data collected.

[0505] Step 2:

[0506] Data preprocessing (server)

[0507] Input: The raw data collected.

[0508] What it does: The server strips HTML tags, normalizes special characters, filters irrelevant data and noise, and generates clean text data.

[0509] Output: Preprocessed and clean text data.

[0510] Step 3:

[0511] Feature extraction (server)

[0512] Input: Preprocessed clean text data.

[0513] Specific operation: The server uses natural language processing techniques (e.g., spaCy or NLTK) to extract features such as keywords, topics, and sentiment analysis results from text data. For video data, it also performs frame analysis and speech-to-text conversion.

[0514] Output: Feature-extracted data (keywords, topics, sentiment scores, etc.).

[0515] Step 4:

[0516] Sentiment analysis (server)

[0517] Input: Feature extracted data.

[0518] What happens: The server uses a sentiment analysis engine (e.g., VADER) to classify specific sentiments (positive, negative, neutral) and calculate a sentiment score.

[0519] Output: Data with sentiment scores.

[0520] Step 5:

[0521] Training and predicting the predictive model (server)

[0522] Input: Preprocessed and sentiment-scored data.

[0523] What it does: The server uses a machine learning model (e.g., Scikit-learn or TensorFlow) to train a model to predict the number of content views and makes the prediction.

[0524] Output: View count prediction results.

[0525] Step 6:

[0526] Generation of prediction results and optimization proposals (server)

[0527] Input: View count prediction results and sentiment score.

[0528] Specific operation: Based on the prediction results, the server generates optimal ad titles and content revision suggestions and provides them to the user.

[0529] Output: Forecast results with optimization suggestions.

[0530] Step 7:

[0531] Content input (terminal)

[0532] Input: Ad content (e.g. title, body, tags).

[0533] Specific actions: The user uses the device to enter data into the content input interface and clicks the submit button.

[0534] Output: The input content data is sent to the server.

[0535] Step 8:

[0536] Sending a prediction request (device)

[0537] Input: The content data entered.

[0538] Specific operation: The terminal converts the content data into JSON format and sends it to the server.

[0539] Output: The content data sent to the server.

[0540] Step 9:

[0541] Check prediction results and optimization suggestions (device)

[0542] Input: Prediction results and optimization suggestions returned by the server.

[0543] Specific operation: The device receives the prediction results and optimization suggestions returned from the server and displays them to the user.

[0544] Output: Prediction results and optimization suggestions displayed to the user.

[0545] Step 10:

[0546] Content optimization and publishing (user)

[0547] Input: Prediction results and optimization suggestions.

[0548] What it does: The user modifies the content based on the displayed predictions and suggestions, then clicks the publish button to publish the optimized content.

[0549] Output: The published optimized content.

[0550] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0551] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0552] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0553] [Second embodiment]

[0554] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0555] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0556] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0557] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0558] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0559] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0560] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0561] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0562] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0563] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0564] In the smart glasses 214, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0565] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0566] Overall system overview

[0567] This invention is a system that collects, preprocesses, and analyzes public data on the Internet and predicts the number of views of content based on the collected data. The system is mainly composed of three entities: a server, a terminal, and a user.

[0568] Server Processing

[0569] The server has the following functions:

[0570] 1. Data Collection

[0571] The server uses web scraping tools and APIs to collect public data from the internet.

[0572] Collections include news sites, blogs, social media posts, and more.

[0573] 2. Data Preprocessing

[0574] The server performs text cleaning on the collected data, specifically removing HTML tags, normalizing special characters, and filtering out irrelevant data.

[0575] The cleaned data is then tokenized and converted into a format suitable for analysis.

[0576] 3. Analysis and feature extraction

[0577] The server uses natural language processing (NLP) techniques to extract features from the text data, including keywords, topics, and sentiment analysis results.

[0578] The server also performs frame analysis and speech-to-text conversion on video data to extract similar features.

[0579] 4. Training the predictive model and making predictions

[0580] The server uses this data to train machine learning models, which use algorithms such as regression analysis and neural networks.

[0581] The trained model is used to predict the number of views for different content.

[0582] 5. Providing prediction results

[0583] The server generates prediction results and provides them to users in real time, including predicted views, diffusion forecasts, and suggestions for optimal presentation and composition.

[0584] Terminal handling

[0585] The terminal provides an interface for users to access the system and input content.

[0586] 1. Content input

[0587] The user enters the title, text, tags, metadata, etc. of the article or video.

[0588] 2. Send a prediction request

[0589] The terminal converts the input content data into JSON format and sends it to the server.

[0590] The prediction results returned from the server are received and displayed to the user.

[0591] User operations

[0592] The user accesses the system through a terminal and optimizes the content in the following steps.

[0593] 1. Content Creation

[0594] The user uses the device to create new article or video content. For example, the user might enter the following as the title: "Predicting the future with AI" and the text: "It is now possible to predict the future using AI technology. In particular, technology that can predict the number of content views with high accuracy is attracting attention."

[0595] 2. Prediction Request

[0596] The user clicks the button to request a prediction of the number of views of the content, and transmits the data to the server.

[0597] The server performs the analysis and predictions and sends the results back to the device.

[0598] 3. Check and correct the prediction results

[0599] The user can check the prediction results displayed on their device, such as "Expected number of views: 1500 PV" or "Recommended title: 'How to read the future with AI technology'."

[0600] The user can then modify the title and body of the content based on the prediction results and publish the optimized content.

[0601] Specific examples

[0602] As a concrete example, we will show a series of steps in which a user creates a blog article, and then uses a server to predict and optimize the number of views of that article.

[0603] 1. A user creates a blog post using a device. The title is "How to predict the future with AI technology" and the body is "It is possible to predict the future using AI technology..."

[0604] 2. The user clicks the "View Count Forecast" button, and the device sends the content data to the server.

[0605] 3. The server analyzes the data and uses a predictive model to predict the number of views.

[0606] 4. The server generates a prediction result and sends it back to the device. For example, "Expected number of views: 1500 PV" and "Recommended title: 'How to predict the future with AI'."

[0607] 5. The user checks the prediction results on their device, changes the title to the recommended title, and publishes the article.

[0608] Through the system of the present invention, users can create effective content based on scientific evidence, which can help create new hit content.

[0609] The processing flow will be explained below.

[0610] Step 1: Data collection

[0611] The server periodically collects public data from the internet using web scraping tools and APIs.

[0612] The server retrieves text and media data from news sites, blogs, social media, etc. and stores it in storage.

[0613] Step 2: Data Preprocessing

[0614] The server removes HTML tags from the collected data, normalizes special characters, and converts them into clean text data.

[0615] The server filters out irrelevant data and noise and converts it into a format suitable for analysis.

[0616] Step 3: Feature extraction

[0617] The server uses natural language processing (NLP) technology to extract features such as keywords, topics, and sentiment analysis results from the text data.

[0618] The server also performs frame analysis and speech-to-text conversion on video data to extract similar features.

[0619] Step 4: Training the predictive model

[0620] The server uses the pre-processed and feature-extracted data to create a training dataset for machine learning.

[0621] The server trains a view prediction model using a regression analysis model or a neural network.

[0622] The server stores the trained model and makes it available for real-time predictions.

[0623] Step 5: Enter content

[0624] The user uses the device interface to input the title, text, tags, metadata, etc. of the article or video.

[0625] Step 6: Send a prediction request

[0626] When the user clicks the "View Count Prediction" button, the device converts the content data into JSON format and sends it to the server.

[0627] Step 7: Data analysis and prediction

[0628] The server analyzes the received content data and performs pre-processing and feature extraction.

[0629] The server uses a trained prediction model to predict the number of views of the content.

[0630] Step 8: Providing prediction results

[0631] The server generates prediction results and returns them to the user, including predicted views, diffusion forecasts, and suggestions for optimal presentation and composition.

[0632] Step 9: Check and correct the predictions

[0633] The user checks the prediction results received on their device, such as "Expected number of views: 1500 PV" and "Recommended title: 'How to read the future with AI technology.'"

[0634] The user can then modify the title and body of the content based on the prediction results and publish the optimized content.

[0635] Example 1

[0636] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0637] In conventional content view prediction systems, the processes from data collection to preprocessing, analysis, prediction, and result provision are often not performed consistently and efficiently. Also, bias toward specific analysis methods or models can make accurate predictions difficult. Another issue is the lack of a means to efficiently transmit content data entered by users to a server in real time and quickly return prediction results. There is a need for a more accurate and efficient content view prediction system that can resolve these issues.

[0638] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0639] In this invention, the server includes means for collecting public data on the Internet, means for preprocessing the collected data, including removing HTML tags, normalizing special characters, and filtering irrelevant data, means for tokenizing the preprocessed data and converting it into a format suitable for analysis, means for extracting features such as keywords, topics, and sentiment analysis results from the text data using natural language processing techniques, means for training a machine learning model based on the analyzed data and predicting the number of views of content, and means for providing the predicted results to users in real time. This makes it possible to efficiently collect data, perform preprocessing and analysis, accurately predict the number of views using the machine learning model, and quickly provide the results to users.

[0640] A "server" is a computer system that collects, analyzes, and predicts public data on a network and provides the results to users.

[0641] "Public data on the internet" is information that is freely accessible online, such as news sites, blogs, and social media posts.

[0642] "Preprocessing" refers to data cleaning tasks such as removing HTML tags from collected data, normalizing special characters, and filtering out irrelevant data.

[0643] "Tokenization" is the process of breaking down text data into units of words and phrases and converting them into a format suitable for analysis.

[0644] "Natural language processing technology" is a technology for extracting features such as keywords, topics, and emotions from text data.

[0645] A "machine learning model" is an algorithm or program that is trained on data to perform a specific task (in this case, predicting content views).

[0646] "Real-time" refers to a form of immediate response where the time between when a user sends data and when they receive the results is very short.

[0647] A "terminal" is a computer or device that allows a user to access the system, enter content, and view results.

[0648] "Content" is a collection of data including article or video titles, text, tags, metadata, etc.

[0649] MODE FOR CARRYING OUT THE INVENTION

[0650] This system collects public data on the Internet, preprocesses and analyzes it, and then predicts the number of views of content based on the collected data. The system is mainly composed of three entities: a server, a terminal, and a user.

[0651] Server Processing

[0652] The server performs data processing and data calculation using the following hardware and software.

[0653] 1. Data Collection

[0654] The server uses web scraping tools (e.g., BeautifulSoup, Scrapy) and APIs (e.g., Twitter API, YouTube Data API) to collect data from news sites, blogs, social media posts, etc.

[0655] Specifically, the server sends an HTTP request to the specified URL and retrieves HTML data or JSON format data.

[0656] 2. Data Preprocessing

[0657] The server performs text cleaning such as removing HTML tags, normalizing special characters, and filtering out irrelevant data. For example, it uses BeautifulSoup to remove HTML tags and regular expressions to normalize special characters.

[0658] The cleaned data is then tokenized and converted into a format suitable for analysis, for example by splitting the text into words using Python's NLTK library.

[0659] 3. Analysis and feature extraction

[0660] The server analyzes the text data using natural language processing (NLP) techniques to extract features such as keywords, topics, and sentiment analysis results. For example, it uses the spaCy library to extract keywords and topics, and calculates sentiment scores using VADER Sentiment Analysis.

[0661] For video data, frame analysis and speech-to-text conversion are performed, for example, by using OpenCV to extract frame-by-frame features and Google Cloud Speech-to-Text to convert the speech to text.

[0662] 4. Training the predictive model and making predictions

[0663] The server trains a machine learning model based on the analyzed feature data, for example, a linear regression model from scikit-learn or a neural network from TensorFlow.

[0664] Use the trained model to predict the number of views for new content.

[0665] 5. Providing prediction results

[0666] The server generates prediction results in JSON format and sends them back to the user's device.

[0667] Terminal handling

[0668] The terminal provides an interface for users to access the system, input content, and check prediction results.

[0669] 1. Content input

[0670] The user uses the device to enter the title, text, tags, and metadata for the article or video.

[0671] 2. Send a prediction request

[0672] The terminal converts the input content data into JSON format and sends it to the server.

[0673] The prediction results returned from the server are received and displayed to the user.

[0674] User operations

[0675] The user accesses the system through a terminal and optimizes the content in the following steps.

[0676] 1. Content Creation

[0677] The user uses the device to create new article or video content. For example, the user might enter the following as the title: "Predicting the future with AI" and the text: "It is now possible to predict the future using AI technology. In particular, technology that can predict the number of content views with high accuracy is attracting attention."

[0678] 2. Prediction Request

[0679] The user clicks the button to request a prediction of the number of views of the content, and the terminal transmits the data to the server.

[0680] 3. Check and correct the prediction results

[0681] The server performs the analysis and predictions and sends the results back to the device. For example, results such as "Expected number of views: 1500 PV" and "Recommended title: 'How to predict the future with AI technology'" are displayed.

[0682] Users can then modify and optimize the title and body of their content based on the prediction results, and then publish the optimized content.

[0683] Through this system, users will be able to create effective content based on scientific evidence, helping to create new hit content.

[0684] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0685] Step 1: Data collection

[0686] Input: URLs of public data on the internet (news sites, blogs, social media posts, etc.)

[0687] How it works: The server sends an HTTP request to a URL to retrieve data in HTML or JSON format, using a web scraping tool (e.g., BeautifulSoup, Scrapy) or an API (e.g., Twitter API, YouTube Data API).

[0688] Output: Collected HTML data and JSON format data

[0689] Step 2: Data Preprocessing - Removing HTML Tags and Normalizing Special Characters

[0690] Input: Collected HTML data or JSON format data

[0691] What it does: The server uses BeautifulSoup to strip HTML tags and regular expressions to normalize special characters.

[0692] Output: Cleaned text data

[0693] Step 3: Data Preprocessing - Data Filtering

[0694] Input: Cleaned text data

[0695] What it does: The server removes data that does not contain certain keywords, for example, filtering out unwanted advertisements or spam posts.

[0696] Output: filtered text data

[0697] Step 4: Tokenization

[0698] Input: filtered text data

[0699] Specific operation: The server uses Python's NLTK library to split the text data into words.

[0700] Output: Tokenized text data

[0701] Step 5: Feature Extraction

[0702] Input: Tokenized text data

[0703] Specific operation: The server extracts features such as keywords, topics, and sentiment scores using natural language processing (NLP) techniques (e.g., spaCy, VADER Sentiment Analysis).

[0704] Output: Feature data such as keywords, topics, and sentiment scores

[0705] Step 6: Analyze the video data

[0706] Input: Video data

[0707] Specific operation: The server uses OpenCV to extract features for each frame and converts the audio into text using Google Cloud Speech-to-Text.

[0708] Output: Frame-by-frame feature data and speech-to-text data

[0709] Step 7: Train the machine learning model

[0710] Input: Feature data (keywords, topics, sentiment scores, etc.) and filtered data

[0711] Specific operation: The server trains a machine learning model based on the feature data using a linear regression model from scikit-learn or a neural network from TensorFlow.

[0712] Output: A trained machine learning model

[0713] Step 8: Predict content views

[0714] Input: New content feature data (keywords, topics, sentiment scores, etc.)

[0715] What it does: The server uses a trained machine learning model to predict the number of views for new content.

[0716] Output: Predicted number of views

[0717] Step 9: Providing prediction results

[0718] Input: Estimated number of views

[0719] Specific operation: The server formats the prediction results in JSON format and sends them to the device.

[0720] Output: Formatted prediction results (JSON format)

[0721] Step 10: Enter content

[0722] Input: Users enter content such as article or video titles, text, tags, and metadata.

[0723] What happens: The device takes these inputs and prepares the data for the prediction request.

[0724] Output: Input content data (JSON format)

[0725] Step 11: Send a prediction request

[0726] Input: Input content data (JSON format)

[0727] Specific operation: The terminal sends content data formatted in JSON format to the server via an HTTP POST request.

[0728] Output: Request sent

[0729] Step 12: Receive and display results

[0730] Input: Prediction results returned from the server (JSON format)

[0731] Specific behavior: The device receives the prediction results and displays them to the user. The device's UI components are used to visually display the results, such as the predicted number of views and recommended titles.

[0732] Output: Prediction results displayed to the user

[0733] This allows users to efficiently collect data, preprocess and analyze it, use machine learning models to accurately predict the number of views, and quickly provide the results to users.

[0734] (Application example 1)

[0735] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0736] While there are many systems that use publicly available data on the Internet to predict content view counts, most of them only provide prediction results and do not include specific optimization suggestions. As a result, users lack the information to effectively optimize titles and thumbnails based on the prediction results, and must rely on trial and error. This makes it difficult to maximize content views.

[0737] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0738] In this invention, the server includes means for collecting public data on the Internet, means for preprocessing and analyzing the collected data, means for predicting the number of views of content based on the preprocessed and analyzed data, means for optimizing the title and thumbnail of the content based on the predicted number of views, and means for providing the prediction results and optimization suggestions to users, thereby enabling users to efficiently and effectively optimize their content and maximize the number of views.

[0739] "Public data on the internet" refers to digital information that is freely accessible from networked sources, such as news sites, blogs, public API data, and social media posts.

[0740] "Methods of collection" refers to the techniques and methods used to obtain data from designated sources, including web scraping and the use of APIs.

[0741] "Preprocessing" refers to a series of operations performed to transform collected data into a format that is easier to analyze, such as text cleaning, normalizing special characters, and filtering out irrelevant data.

[0742] "Analytical means" refers to techniques and methods for analyzing collected and pre-processed data, including the application of natural language processing techniques and machine learning algorithms.

[0743] "Methods for predicting views" refers to techniques or methods for predicting the number of views of content based on analyzed data, including models such as regression analysis and neural networks.

[0744] "Optimization means" refers to technologies and methods for effectively changing or suggesting content titles and thumbnails based on the prediction results.

[0745] "Means of providing" refers to the interface and functions for notifying and displaying prediction results and optimization suggestions to users, including the user interface and notification functions.

[0746] This system effectively maximizes the number of content views by collecting, preprocessing, and analyzing public data on the Internet and providing users with prediction results and optimization suggestions. This system is mainly composed of three entities: a server, a terminal, and a user.

[0747] Server Processing

[0748] The server has the following functions:

[0749] 1. Data Collection:

[0750] The server uses web scraping tools and APIs to collect public data from the internet, such as news sites, blogs, and social media posts.

[0751] 2. Data Preprocessing:

[0752] The server performs text cleaning on the collected data, specifically removing HTML tags, normalizing special characters, and filtering out irrelevant data.

[0753] 3. Analysis and feature extraction:

[0754] The server uses natural language processing (NLP) techniques to extract features from the text data, including keywords, topics, and sentiment analysis results. It also uses the Python library Scikit-learn to vectorize the text data using TF-IDF.

[0755] 4. Views forecast:

[0756] The server trains a machine learning model based on the pre-processed and feature-extracted data to predict the number of views. It uses algorithms such as regression analysis and neural networks to build the predictive model.

[0757] 5. Optimization:

[0758] Based on the prediction results, the server generates suggestions to optimize the title and thumbnail of the content.

[0759] 6. Providing prediction results:

[0760] The server provides the prediction results and optimization suggestions to the user in real time.

[0761] Terminal handling

[0762] The terminal provides an interface for users to access the system and input content.

[0763] 1. Content Input:

[0764] Users use the device to input the title, text, tags, metadata, etc. of articles and videos.

[0765] 2. Send a prediction request:

[0766] The terminal converts the input content data into JSON format and sends it to the server.

[0767] 3. Receiving and displaying prediction results:

[0768] The terminal receives the prediction results and optimization suggestions returned from the server and displays them to the user.

[0769] User operations

[0770] The user accesses the system through a terminal and optimizes the content in the following steps.

[0771] 1. Content Creation:

[0772] Users use their devices to create new article or video content, for example by typing: "How to read the future with AI technology."

[0773] 2. Prediction request:

[0774] The user clicks the button requesting a prediction of the number of views of the content, and transmits data from the terminal to the server.

[0775] 3. Check and correct the predictions:

[0776] Users can check the prediction results and optimization suggestions displayed on their devices. For example, they may see "Expected page views: 1,500 PV" or "Recommended title: 'How to predict the future with AI technology.'" Based on this, users can revise the title and thumbnail and publish the optimized content.

[0777] Example prompt sentence:

[0778] The system collects and preprocesses data in the following way, then analyzes it to predict the number of views of the content. The specific steps are as follows:

[0779] 1. Collect news sites, blogs, and social media posts using web scraping tools and APIs.

[0780] 2. Clean the collected data and convert it into a suitable format for analysis.

[0781] 3. Extract features using natural language processing techniques.

[0782] 4. Train a machine learning model to predict viewership.

[0783] 5. Provide users with real-time prediction results.

[0784] Users can generate optimal titles and descriptions for various content to maximize views. For example, a video title like "How to Read the Future with AI Technology" is expected to attract many viewers according to the predictive model.

[0785] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0786] Step 1: Data collection

[0787] The server collects public data from news sites, blogs, social media posts, and other sources on the internet. To do this, it uses a web scraping tool or API. Specifically, it sends HTTP requests to specified URLs and parses the resulting HTML data to extract usable information. A list of URLs is given as input, and text data is obtained as output.

[0788] Step 2: Data Preprocessing

[0789] The server preprocesses the collected text data by removing HTML tags, normalizing special characters, and filtering irrelevant data. Raw text data is given as input, and clean text data is obtained as output. Regular expressions and text cleaning libraries are used in this process.

[0790] Step 3: Feature extraction

[0791] The server extracts features using the preprocessed text data. To do this, it uses natural language processing (NLP) techniques. Specifically, it uses the Python library Scikit-learn to perform TF-IDF vectorization. Clean text data is given as input, and vectorized feature data is obtained as output.

[0792] Step 4: Train the machine learning model

[0793] The server uses the feature data to train a machine learning model. This uses algorithms such as regression analysis and neural networks. Specifically, it uses the LinearRegression model from Scikit-learn. Vectorized feature data and label data (past view counts) are given as input, and the trained model is obtained as output.

[0794] Step 5: Predicting views

[0795] The server uses the trained machine learning model to predict the number of views for new content data provided by the user. The server receives feature data of the new content as input and obtains the predicted number of views as output.

[0796] Step 6: Optimization suggestions

[0797] The server generates optimal title and thumbnail suggestions based on the prediction results, taking into account past data on similar methods that have generated large numbers of views. The predicted number of views and content feature data are given as input, and the output is an optimized suggestion.

[0798] Step 7: Providing prediction results and optimization suggestions

[0799] The server provides the prediction results and optimization suggestions to the user in real time. This information is displayed to the user through the terminal. The prediction results and optimization suggestions are given as input and displayed to the user as output.

[0800] Step 8: Optimize your content for users

[0801] The user modifies the title and thumbnail of the content based on the provided prediction results and optimization suggestions, which effectively maximizes the number of views of the content. The prediction results and optimization suggestions are given as input, and the optimized content is obtained as output.

[0802] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0803] Overall system overview

[0804] This system collects, preprocesses, and analyzes public data on the Internet, and then predicts the number of views of content based on that data. Furthermore, it incorporates an emotion engine that recognizes user emotions, allowing for predictions that take emotion data into account. This system is primarily comprised of three components: a server, a terminal, and a user.

[0805] Server Processing

[0806] The server has the following functions:

[0807] 1. Data Collection

[0808] The server periodically collects public data from the internet using web scraping tools and APIs.

[0809] The server retrieves text and media data from news sites, blogs, social media, etc. and stores it in storage.

[0810] 2. Data Preprocessing

[0811] The server removes HTML tags from the collected data, normalizes special characters, and converts them into clean text data.

[0812] The server filters out irrelevant data and noise and converts it into a format suitable for analysis.

[0813] 3. Feature Extraction

[0814] The server uses natural language processing (NLP) technology to extract features such as keywords, topics, and sentiment analysis results from the text data.

[0815] The server also performs frame analysis and speech-to-text conversion on video data to extract similar features.

[0816] 4. Analysis by Emotion Engine

[0817] The server uses an emotion engine to recognize emotions from text data about users and content, for example classifying positive, negative, and neutral emotions.

[0818] The server incorporates the emotion data into the pre-processing and feature extraction data.

[0819] 5. Training the predictive model and making predictions

[0820] The server uses the pre-processed and feature-extracted data and the sentiment data to create a training dataset for machine learning.

[0821] The server trains a view prediction model using a regression analysis model or a neural network.

[0822] The server stores the trained model and makes it available for real-time predictions.

[0823] 6. Providing prediction results

[0824] The server generates prediction results and provides them to users in real time, including predicted views, virality predictions, suggestions for optimal presentation and composition, and sentiment-based recommendations.

[0825] Terminal handling

[0826] The terminal provides an interface for users to access the system and input content.

[0827] 1. Content input

[0828] The user enters the title, text, tags, metadata, etc. of the article or video.

[0829] 2. Send a prediction request

[0830] The terminal converts the input content data into JSON format and sends it to the server.

[0831] 3. Check the prediction results

[0832] The terminal receives the prediction results returned from the server and displays them to the user.

[0833] User operations

[0834] The user accesses the system through a terminal and optimizes the content in the following steps.

[0835] 1. Content Creation

[0836] The user uses the device to create new article or video content. For example, the user might enter the following as the title: "Predicting the future with AI" and the text: "It is now possible to predict the future using AI technology. In particular, technology that can predict the number of content views with high accuracy is attracting attention."

[0837] 2. Prediction Request

[0838] The user clicks the button to request a prediction of the number of views of the content, and transmits the data to the server.

[0839] The server performs the analysis and predictions and sends the results back to the device.

[0840] 3. Check and correct the prediction results

[0841] The user checks the prediction results displayed on their device, such as "Expected number of views: 1500 PV," "Recommended title: 'How to read the future with AI technology,'" and "Emotion score: High."

[0842] The user can then modify the title and body of the content based on the prediction results and publish the optimized content.

[0843] Specific examples

[0844] As a concrete example, we show a series of steps in which a user creates a blog post, a server is used to predict the number of views of the post, and optimization is performed taking into account emotional data.

[0845] 1. A user creates a blog post using a device. The title is "How to predict the future with AI technology" and the body is "It is possible to predict the future using AI technology..."

[0846] 2. The user clicks the "View Count Forecast" button, and the device sends the content data to the server.

[0847] 3. The server analyzes the data and extracts emotion data using an emotion engine.

[0848] 4. The server uses the trained prediction model to predict the number of views and generates a prediction result.

[0849] 5. The server sends the prediction results back to the device. For example, "Expected number of views: 1500 PV", "Recommended title: 'How to predict the future with AI'", "Sentiment score: High".

[0850] 6. The user checks the prediction results on their device, changes the title to the recommended title, and publishes the optimized article.

[0851] Through the system of the present invention, users can create effective content based on scientific evidence that takes into account emotional data, thereby supporting the creation of new hit content.

[0852] The processing flow will be explained below.

[0853] Step 1: Data collection

[0854] The server periodically collects public data from the internet using web scraping tools and APIs.

[0855] The server retrieves text and media data from news sites, blogs, social media, etc. and stores it in storage.

[0856] Step 2: Data Preprocessing

[0857] The server removes HTML tags from the collected data, normalizes special characters, and converts them into clean text data.

[0858] The server filters out irrelevant data and noise and converts it into a format suitable for analysis.

[0859] Step 3: Feature extraction

[0860] The server uses natural language processing (NLP) technology to extract features such as keywords, topics, and sentiment analysis results from the text data.

[0861] The server also performs frame analysis and speech-to-text conversion on video data to extract similar features.

[0862] Step 4: Analysis by Emotion Engine

[0863] The server uses an emotion engine to recognize emotions from text data about users and content.

[0864] The server classifies and scores positive, negative, and neutral sentiment within the text data.

[0865] Step 5: Train the predictive model

[0866] The server uses the pre-processed and feature-extracted data and the sentiment data to create a training dataset for machine learning.

[0867] The server trains a view prediction model using a regression analysis model or a neural network.

[0868] The server stores the trained model and makes it available for real-time predictions.

[0869] Step 6: Enter content

[0870] The user uses the device interface to input the title, text, tags, metadata, etc. of the article or video.

[0871] Step 7: Send a prediction request

[0872] When the user clicks the "View Count Prediction" button, the device converts the content data into JSON format and sends it to the server.

[0873] Step 8: Data analysis and prediction

[0874] The server analyzes the received content data and performs pre-processing and feature extraction.

[0875] The server analyzes the emotion of the content using an emotion engine and calculates an emotion score.

[0876] The server uses a trained prediction model to predict the number of views of the content.

[0877] Step 9: Providing prediction results

[0878] The server generates and returns prediction results to the user, including expected views, virality predictions, suggestions for optimal wording and composition, and sentiment scores.

[0879] Step 10: Check and correct the prediction results

[0880] The user checks the prediction results received on their device, which may include, for example, "Expected number of views: 1500 PV," "Recommended title: 'How to read the future with AI technology,'" and "Emotion score: High."

[0881] The user can then modify the title and body of the content based on the prediction results.

[0882] The user publishes the optimized content.

[0883] Example 2

[0884] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0885] Conventional systems that use public data on the Internet to predict the number of views of content are limited to preprocessing and analysis of collected data, and it is difficult to make predictions that take user emotions into account. As a result, prediction accuracy is low and it is difficult to create effective content.

[0886] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for collecting public data on the Internet, means for preprocessing and analyzing the collected data, means for predicting the number of views of content based on the preprocessed and analyzed data, means for extracting and analyzing emotion data using an emotion engine that recognizes the user's emotions, and means for providing the prediction result to the user. This makes it possible to predict the number of views of content with high accuracy, taking the user's emotions into consideration.

[0887] "Public data on the Internet" refers to all information, including text data and media data, that is publicly available on the Internet.

[0888] "Means of collection" refers to the function of regularly obtaining publicly available data on the Internet using web scraping tools and APIs.

[0889] "Preprocessing" refers to the process of removing HTML tags and normalizing special characters from collected data to convert it into clean text data.

[0890] "Means of analysis" refers to the function of analyzing preprocessed data using natural language processing technology, etc., and extracting necessary features and information.

[0891] "Means for predicting the number of content views" refers to a function that uses a machine learning model to predict the number of content views based on pre-processed and analyzed data.

[0892] An "emotion engine" is software or a service for analyzing user emotions, and has the ability to recognize positive, negative, and neutral emotions from text data.

[0893] "Emotion data" is the result data of emotions related to users and content extracted by the emotion engine.

[0894] The "means for providing predicted results" refers to the function of the server to generate predicted results and provide them to the user in real time.

[0895] "Natural language processing technology" is a technology for analyzing text data and extracting features, and includes keyword extraction, topic modeling, sentiment analysis, etc.

[0896] "Content data entered by the user" refers to information such as the title, text, tags, and metadata of articles and videos that the user sends to the system via their terminal.

[0897] "Server" refers to a computer system that centrally handles functions such as data collection, preprocessing, analysis, emotion recognition, and view count prediction.

[0898] "Terminal" refers to the device or application that a User uses to access the System and input Content.

[0899] This system collects, preprocesses, and analyzes public data on the Internet, and predicts the number of views of content based on that data. Furthermore, it combines an emotion engine that recognizes user emotions, allowing for predictions that take emotion data into account.

[0900] Server functions and processing contents

[0901] 1. Data Collection

[0902] The server uses web scraping tools (e.g., BeautifulSoup, Scrapy) and APIs to collect public data from news sites, blogs, social media, etc. For example, article text from news sites and social media posts are stored in a database. The server periodically collects this data and stores it in storage.

[0903] 2. Data Preprocessing

[0904] The server converts the collected data into clean text data by removing HTML tags and normalizing special characters. Specifically, it uses BeautifulSoup to remove unnecessary tags and format the data in a unified format. It also filters out advertisements and irrelevant parts to create a dataset suitable for analysis.

[0905] 3. Feature Extraction

[0906] The server uses natural language processing (NLP) techniques (e.g., spaCy, NLTK) to extract features such as keywords, topics, and sentiment analysis results from text data. For example, it performs topic modeling and sentiment scoring. For video data, it performs frame analysis using OpenCV and speech-to-text conversion using Google Speech-to-Text to extract similar features.

[0907] 4. Analysis by Emotion Engine

[0908] The server analyzes the user's emotions using an emotion engine (e.g., Natural Language Understanding), which classifies positive, negative, and neutral emotions from the text data and integrates the emotion data into a feature dataset.

[0909] 5. Training the predictive model and making predictions

[0910] The server uses the preprocessed and feature-extracted dataset to train a machine learning model (e.g., TensorFlow, PyTorch). The trained predictive model is used to predict the number of views of new content in real time.

[0911] 6. Providing prediction results

[0912] The server uses the trained model to generate predictions and provide them to the user, such as generating JSON data containing predicted views and sentiment-based recommendations, and sending it back to the device.

[0913] User and Device Capabilities

[0914] 1. Content input

[0915] The user uses the device to enter the title, text, tags, and metadata of a new article or video. For example, the title might be "Predicting the future with AI" and the text might be "AI technology makes it possible to predict the future...".

[0916] 2. Send a prediction request

[0917] When the user clicks the "View Count Prediction" button, the device converts the entered content data into JSON format and sends it to the server.

[0918] 3. Check the prediction results

[0919] The device receives the prediction results sent from the server and displays them to the user, such as "Expected number of views: 1500 PV," a recommended title "How to predict the future with AI technology," and an emotion score of "High."

[0920] 4. Modify and optimize your content

[0921] The user checks the prediction results and modifies the content as necessary, for example, changing the title to a recommended title or adjusting the wording of the body text. The optimized content is then published by the user.

[0922] Examples and prompts

[0923] As a concrete example, we show a series of steps in which a user creates a blog post, a server is used to predict the number of views of the post, and optimization is performed taking into account emotional data.

[0924] Title: "How to predict the future with AI technology"

[0925] Body text: "AI technology is making it possible to predict the future. In particular, technology that can predict the number of content views with high accuracy is attracting attention."

[0926] An example prompt might be, "What is the expected number of views for this article?"

[0927] Through this system, users will be able to create effective content based on scientific evidence, helping to create new hit content.

[0928] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0929] Step 1: Data collection

[0930] Input: The server retrieves public data from the internet (e.g., news articles, blog posts, social media posts, etc.) from a specified URL or API.

[0931] What it does: The server collects public data using web scraping tools (e.g., BeautifulSoup, Scrapy) or APIs. For example, it scrapes articles from news sites with BeautifulSoup and retrieves social media posts through APIs.

[0932] Output: The collected raw data is stored in the server's storage. It includes files in various formats, such as text data, image data, and audio data.

[0933] Step 2: Data Preprocessing

[0934] Input: Raw data collected by the server.

[0935] What it does: The server uses BeautifulSoup to strip HTML tags, normalize special characters, and perform noise filtering to remove irrelevant data, such as removing ads and script tags from HTML and resolving encoding issues.

[0936] Output: Clean text data and structured data. For example, ads and irrelevant information are removed and a normalized text file is output.

[0937] Step 3: Feature extraction

[0938] Input: Clean text data.

[0939] Specific operation: The server uses natural language processing (NLP) techniques (e.g., spaCy, NLTK) to extract features such as keywords, topics, and sentiment analysis results. For video data, it performs frame analysis using OpenCV and converts the audio to text using Google Speech-to-Text.

[0940] Output: A dataset of extracted features, including a keyword list and sentiment scores from text data, and important scenes extracted from video frames.

[0941] Step 4: Analysis by Emotion Engine

[0942] Input: Extracted feature dataset.

[0943] Specific operation: The server uses an emotion engine (e.g., Natural Language Understanding) to classify the emotion of each piece of text data and calculate an emotion score, for example, classifying it as "positive," "negative," or "neutral."

[0944] Output: A feature dataset with sentiment data added, e.g., sentiment scores for each text data.

[0945] Step 5: Train the predictive model and make predictions

[0946] Input: Preprocessed and feature extracted dataset.

[0947] What it does: The server uses machine learning models (e.g., TensorFlow, PyTorch) to train regression analysis or neural network models using this data, and then uses the trained models to predict the number of views for new content.

[0948] Output: The trained prediction model and prediction results, such as predicted views and suggested titles for new articles.

[0949] Step 6: Providing prediction results

[0950] Input: JSON data containing prediction results.

[0951] Specific operation: The server generates a prediction result and sends it back to the device. For example, "Expected number of views: 1500 PV" or "Recommended title: 'How to predict the future with AI technology'".

[0952] Output: The prediction result is sent to the device so that the user can check it.

[0953] Step 7: User input and prediction request

[0954] Input: User-generated content data (article title, body, tags, metadata, etc.).

[0955] Specific operation: The user inputs content through the terminal and clicks the "View Count Prediction" button. The terminal converts the content data into JSON format and sends it to the server.

[0956] Output: The content data sent to the server.

[0957] Step 8: Check the prediction results and modify the content

[0958] Input: Prediction results sent from the server.

[0959] Specific operation: The user checks the prediction results on the device and modifies the content as necessary. For example, the title can be changed to "How to read the future with AI technology" based on the prediction results.

[0960] Output: The optimized content is published.

[0961] As described above, data input, data processing, data calculation, and output are performed in each processing step, enabling users to create effective content by utilizing highly accurate prediction results and emotion data.

[0962] (Application example 2)

[0963] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0964] Conventional content view count prediction systems simply predicted the number of views without considering user emotions, making it difficult to optimize content more effectively. In particular, in the advertising field, where emotion-based optimization is required, there was a lack of functionality to suggest optimal titles and content.

[0965] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0966] In this invention, the server includes means for collecting public data on the Internet, means for preprocessing and analyzing the collected data, means for predicting the number of views of content based on the preprocessed and analyzed data, means for recognizing user emotions using a sentiment analysis engine and making predictions taking the emotional data into consideration, means for providing the prediction results and optimization suggestions to the user, and means for generating optimal titles and content revision suggestions for advertising content, thereby enabling the prediction and optimization of the number of views of advertising content that reflects user emotions.

[0967] "Public data on the internet" refers to information that is freely accessible on websites, blogs, social media platforms, etc.

[0968] "Means of collection" refers to the process of using web scraping tools or APIs to obtain public data on a server.

[0969] "Preprocessing" refers to the process of removing unnecessary HTML tags from collected data, normalizing special characters, and converting the data into a format suitable for analysis.

[0970] "Means of analysis" refers to the process of analyzing data using natural language processing techniques and extracting features such as keywords, topics, and sentiment.

[0971] "Method for predicting content views" refers to the process of using a machine learning model to predict content views based on pre-processed and analyzed data.

[0972] "Sentiment analysis engine" refers to an algorithm that recognizes positive, negative, and neutral emotions from text data entered by the user.

[0973] "Means for making predictions taking into account emotional data" refers to a process of incorporating emotional data into a view count prediction model to make view count predictions that take into account the influence of emotions.

[0974] "Means for providing prediction results and optimization suggestions to users" refers to a process for providing users with prediction results as well as suggestions for more effective content expression and configuration.

[0975] "Means for generating optimal titles and suggested content modifications for advertising content" refers to the process of generating effective titles and suggested text modifications for advertising content based on prediction results and emotional data.

[0976] The present invention relates to a system that collects, preprocesses, and analyzes public data on the Internet and predicts the number of views of content based on that data. This system, combined with a sentiment analysis engine, performs advanced predictions that take emotional data into account and provides optimal titles and content revisions, particularly in the advertising field. A specific embodiment of this system is described below.

[0977] Server Processing

[0978] The server has the following main functions:

[0979] 1. Data Collection

[0980] The server periodically collects public data from the internet using web scraping tools and APIs.

[0981] For example, text and media data is acquired from news sites, blogs, and social media and stored in a database.

[0982] 2. Data Preprocessing

[0983] The server removes HTML tags from the collected data, normalizes special characters, and converts them into clean text data.

[0984] It also filters out irrelevant data and noise and converts it into a format suitable for analysis.

[0985] 3. Feature Extraction

[0986] The server uses natural language processing technology (e.g., spaCy or NLTK) to extract features such as keywords, topics, and sentiment analysis results from the text data.

[0987] In the case of video data, frame analysis and speech-to-text conversion are performed to extract similar features.

[0988] 4. Analysis using a sentiment analysis engine

[0989] The server uses a sentiment analysis engine (such as VADER) to classify the user's sentiment from the text data.

[0990] The classified sentiment data (positive, negative, neutral) is incorporated into the preprocessing and feature extraction data.

[0991] 5. Training the predictive model and making predictions

[0992] The server uses the preprocessed and feature-extracted data and sentiment data to train a machine learning model (e.g., Scikit-learn or TensorFlow).

[0993] The trained model is used to predict the number of views of the content.

[0994] 6. Providing prediction results and optimization suggestions

[0995] The server generates and provides the user with optimal representation and composition suggestions along with the prediction results.

[0996] Specifically, it generates optimal titles and suggested content revisions for advertising content, reflecting user emotional data.

[0997] Terminal handling

[0998] The terminal provides an interface for users to access the system and input content.

[0999] 1. Content input

[1000] Users enter the title, text, tags, metadata, etc. of the article or video.

[1001] 2. Send a prediction request

[1002] The terminal converts the input content data into JSON format and sends it to the server.

[1003] 3. Check the prediction results and optimization suggestions

[1004] The terminal receives the prediction results and optimization suggestions returned from the server and displays them to the user.

[1005] User operations

[1006] A user accesses the system through a terminal and optimizes advertising content by following the steps below.

[1007] 1. Content Creation

[1008] The user uses the device to create new advertising content, for example, by typing "We've launched a new smartwatch. Learn more!"

[1009] 2. Prediction Request

[1010] The user clicks the button to request a prediction of the number of views of the content, and transmits the data to the server.

[1011] 3. Check and modify the prediction results and optimization suggestions

[1012] Users can view the prediction results and optimization suggestions displayed on their device, such as "Expected page views: 3,000 PV," "Recommended title: 'Limited sale! Get it now,'" and "Sentiment score: High."

[1013] Users optimize and publish content based on prediction results and suggestions.

[1014] Specific examples

[1015] When a user submits an ad to the system saying, "We've released a new smartwatch. Click here for details!", the system performs sentiment analysis and predicts the number of views. The predicted results are displayed as "Sentiment score: Medium," "Expected number of views: 2000 PV," and "Recommended title: 'Limited time only! Check out the details of the limited edition smartwatch.'" The user can then change the displayed title to optimize the ad.

[1016] Prompt Sentence Examples

[1017] Prompt statement:

[1018] The ad data entered by the user is listed below. Please perform sentiment analysis and respond with a view prediction and optimal ad modification suggestions.

[1019] Ad title: New smartwatch released

[1020] Ad text: We've launched a new smartwatch. Learn more here!

[1021] Tags: smartwatch sale

[1022] Expected output:

[1023] Estimated number of views: XXXX PV

[1024] Recommended title: 'Now only! Check out the details of the limited edition smartwatch'

[1025] Sentiment score: High / Medium / Low

[1026] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1027] Step 1:

[1028] Data collection (server)

[1029] Input: Public data from the internet (e.g., news articles, blog posts, social media posts).

[1030] What it does: The server periodically collects this public data using web scraping tools and APIs and stores it in a database.

[1031] Output: The raw data collected.

[1032] Step 2:

[1033] Data preprocessing (server)

[1034] Input: The raw data collected.

[1035] What it does: The server strips HTML tags, normalizes special characters, filters irrelevant data and noise, and generates clean text data.

[1036] Output: Preprocessed and clean text data.

[1037] Step 3:

[1038] Feature extraction (server)

[1039] Input: Preprocessed clean text data.

[1040] Specific operation: The server uses natural language processing techniques (e.g., spaCy or NLTK) to extract features such as keywords, topics, and sentiment analysis results from text data. For video data, it also performs frame analysis and speech-to-text conversion.

[1041] Output: Feature-extracted data (keywords, topics, sentiment scores, etc.).

[1042] Step 4:

[1043] Sentiment analysis (server)

[1044] Input: Feature extracted data.

[1045] What happens: The server uses a sentiment analysis engine (e.g., VADER) to classify specific sentiments (positive, negative, neutral) and calculate a sentiment score.

[1046] Output: Data with sentiment scores.

[1047] Step 5:

[1048] Training and predicting the predictive model (server)

[1049] Input: Preprocessed and sentiment-scored data.

[1050] What it does: The server uses a machine learning model (e.g., Scikit-learn or TensorFlow) to train a model to predict the number of content views and makes the prediction.

[1051] Output: View count prediction results.

[1052] Step 6:

[1053] Generation of prediction results and optimization proposals (server)

[1054] Input: View count prediction results and sentiment score.

[1055] Specific operation: Based on the prediction results, the server generates optimal ad titles and content revision suggestions and provides them to the user.

[1056] Output: Forecast results with optimization suggestions.

[1057] Step 7:

[1058] Content input (terminal)

[1059] Input: Ad content (e.g. title, body, tags).

[1060] Specific actions: The user uses the device to enter data into the content input interface and clicks the submit button.

[1061] Output: The input content data is sent to the server.

[1062] Step 8:

[1063] Sending a prediction request (device)

[1064] Input: The content data entered.

[1065] Specific operation: The terminal converts the content data into JSON format and sends it to the server.

[1066] Output: The content data sent to the server.

[1067] Step 9:

[1068] Check prediction results and optimization suggestions (device)

[1069] Input: Prediction results and optimization suggestions returned by the server.

[1070] Specific operation: The device receives the prediction results and optimization suggestions returned from the server and displays them to the user.

[1071] Output: Prediction results and optimization suggestions displayed to the user.

[1072] Step 10:

[1073] Content optimization and publishing (user)

[1074] Input: Prediction results and optimization suggestions.

[1075] What it does: The user modifies the content based on the displayed predictions and suggestions, then clicks the publish button to publish the optimized content.

[1076] Output: The published optimized content.

[1077] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1078] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1079] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[1080] [Third embodiment]

[1081] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1082] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[1083] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1084] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[1085] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1086] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1087] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1088] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1089] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1090] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1091] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1092] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[1093] Overall system overview

[1094] This invention is a system that collects, preprocesses, and analyzes public data on the Internet and predicts the number of views of content based on the collected data. The system is mainly composed of three entities: a server, a terminal, and a user.

[1095] Server Processing

[1096] The server has the following functions:

[1097] 1. Data Collection

[1098] The server uses web scraping tools and APIs to collect public data from the internet.

[1099] Collections include news sites, blogs, social media posts, and more.

[1100] 2. Data Preprocessing

[1101] The server performs text cleaning on the collected data, specifically removing HTML tags, normalizing special characters, and filtering out irrelevant data.

[1102] The cleaned data is then tokenized and converted into a format suitable for analysis.

[1103] 3. Analysis and feature extraction

[1104] The server uses natural language processing (NLP) techniques to extract features from the text data, including keywords, topics, and sentiment analysis results.

[1105] The server also performs frame analysis and speech-to-text conversion on video data to extract similar features.

[1106] 4. Training the predictive model and making predictions

[1107] The server uses this data to train machine learning models, which use algorithms such as regression analysis and neural networks.

[1108] The trained model is used to predict the number of views for different content.

[1109] 5. Providing prediction results

[1110] The server generates prediction results and provides them to users in real time, including predicted views, diffusion forecasts, and suggestions for optimal presentation and composition.

[1111] Terminal handling

[1112] The terminal provides an interface for users to access the system and input content.

[1113] 1. Content input

[1114] The user enters the title, text, tags, metadata, etc. of the article or video.

[1115] 2. Send a prediction request

[1116] The terminal converts the input content data into JSON format and sends it to the server.

[1117] The prediction results returned from the server are received and displayed to the user.

[1118] User operations

[1119] The user accesses the system through a terminal and optimizes the content in the following steps.

[1120] 1. Content Creation

[1121] The user uses the device to create new article or video content. For example, the user might enter the following as the title: "Predicting the future with AI" and the text: "It is now possible to predict the future using AI technology. In particular, technology that can predict the number of content views with high accuracy is attracting attention."

[1122] 2. Prediction Request

[1123] The user clicks the button to request a prediction of the number of views of the content, and transmits the data to the server.

[1124] The server performs the analysis and predictions and sends the results back to the device.

[1125] 3. Check and correct the prediction results

[1126] The user can check the prediction results displayed on their device, such as "Expected number of views: 1500 PV" or "Recommended title: 'How to read the future with AI technology'."

[1127] The user can then modify the title and body of the content based on the prediction results and publish the optimized content.

[1128] Specific examples

[1129] As a concrete example, we will show a series of steps in which a user creates a blog article, and then uses a server to predict and optimize the number of views of that article.

[1130] 1. A user creates a blog post using a device. The title is "How to predict the future with AI technology" and the body is "It is possible to predict the future using AI technology..."

[1131] 2. The user clicks the "View Count Forecast" button, and the device sends the content data to the server.

[1132] 3. The server analyzes the data and uses a predictive model to predict the number of views.

[1133] 4. The server generates a prediction result and sends it back to the device. For example, "Expected number of views: 1500 PV" and "Recommended title: 'How to predict the future with AI'."

[1134] 5. The user checks the prediction results on their device, changes the title to the recommended title, and publishes the article.

[1135] Through the system of the present invention, users can create effective content based on scientific evidence, which can help create new hit content.

[1136] The processing flow will be explained below.

[1137] Step 1: Data collection

[1138] The server periodically collects public data from the internet using web scraping tools and APIs.

[1139] The server retrieves text and media data from news sites, blogs, social media, etc. and stores it in storage.

[1140] Step 2: Data Preprocessing

[1141] The server removes HTML tags from the collected data, normalizes special characters, and converts them into clean text data.

[1142] The server filters out irrelevant data and noise and converts it into a format suitable for analysis.

[1143] Step 3: Feature extraction

[1144] The server uses natural language processing (NLP) technology to extract features such as keywords, topics, and sentiment analysis results from the text data.

[1145] The server also performs frame analysis and speech-to-text conversion on video data to extract similar features.

[1146] Step 4: Training the predictive model

[1147] The server uses the pre-processed and feature-extracted data to create a training dataset for machine learning.

[1148] The server trains a view prediction model using a regression analysis model or a neural network.

[1149] The server stores the trained model and makes it available for real-time predictions.

[1150] Step 5: Enter content

[1151] The user uses the device interface to input the title, text, tags, metadata, etc. of the article or video.

[1152] Step 6: Send a prediction request

[1153] When the user clicks the "View Count Prediction" button, the device converts the content data into JSON format and sends it to the server.

[1154] Step 7: Data analysis and prediction

[1155] The server analyzes the received content data and performs pre-processing and feature extraction.

[1156] The server uses a trained prediction model to predict the number of views of the content.

[1157] Step 8: Providing prediction results

[1158] The server generates prediction results and returns them to the user, including predicted views, diffusion forecasts, and suggestions for optimal presentation and composition.

[1159] Step 9: Check and correct the predictions

[1160] The user checks the prediction results received on their device, such as "Expected number of views: 1500 PV" and "Recommended title: 'How to read the future with AI technology.'"

[1161] The user can then modify the title and body of the content based on the prediction results and publish the optimized content.

[1162] Example 1

[1163] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1164] In conventional content view prediction systems, the processes from data collection to preprocessing, analysis, prediction, and result provision are often not performed consistently and efficiently. Also, bias toward specific analysis methods or models can make accurate predictions difficult. Another issue is the lack of a means to efficiently transmit content data entered by users to a server in real time and quickly return prediction results. There is a need for a more accurate and efficient content view prediction system that can resolve these issues.

[1165] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1166] In this invention, the server includes means for collecting public data on the Internet, means for preprocessing the collected data, including removing HTML tags, normalizing special characters, and filtering irrelevant data, means for tokenizing the preprocessed data and converting it into a format suitable for analysis, means for extracting features such as keywords, topics, and sentiment analysis results from the text data using natural language processing techniques, means for training a machine learning model based on the analyzed data and predicting the number of views of content, and means for providing the predicted results to users in real time. This makes it possible to efficiently collect data, perform preprocessing and analysis, accurately predict the number of views using the machine learning model, and quickly provide the results to users.

[1167] A "server" is a computer system that collects, analyzes, and predicts public data on a network and provides the results to users.

[1168] "Public data on the internet" is information that is freely accessible online, such as news sites, blogs, and social media posts.

[1169] "Preprocessing" refers to data cleaning tasks such as removing HTML tags from collected data, normalizing special characters, and filtering out irrelevant data.

[1170] "Tokenization" is the process of breaking down text data into units of words and phrases and converting them into a format suitable for analysis.

[1171] "Natural language processing technology" is a technology for extracting features such as keywords, topics, and emotions from text data.

[1172] A "machine learning model" is an algorithm or program that is trained on data to perform a specific task (in this case, predicting content views).

[1173] "Real-time" refers to a form of immediate response where the time between when a user sends data and when they receive the results is very short.

[1174] A "terminal" is a computer or device that allows a user to access the system, enter content, and view results.

[1175] "Content" is a collection of data including article or video titles, text, tags, metadata, etc.

[1176] MODE FOR CARRYING OUT THE INVENTION

[1177] This system collects public data on the Internet, preprocesses and analyzes it, and then predicts the number of views of content based on the collected data. The system is mainly composed of three entities: a server, a terminal, and a user.

[1178] Server Processing

[1179] The server performs data processing and data calculation using the following hardware and software.

[1180] 1. Data Collection

[1181] The server uses web scraping tools (e.g., BeautifulSoup, Scrapy) and APIs (e.g., Twitter API, YouTube Data API) to collect data from news sites, blogs, social media posts, etc.

[1182] Specifically, the server sends an HTTP request to the specified URL and retrieves HTML data or JSON format data.

[1183] 2. Data Preprocessing

[1184] The server performs text cleaning such as removing HTML tags, normalizing special characters, and filtering out irrelevant data. For example, it uses BeautifulSoup to remove HTML tags and regular expressions to normalize special characters.

[1185] The cleaned data is then tokenized and converted into a format suitable for analysis, for example by splitting the text into words using Python's NLTK library.

[1186] 3. Analysis and feature extraction

[1187] The server analyzes the text data using natural language processing (NLP) techniques to extract features such as keywords, topics, and sentiment analysis results. For example, it uses the spaCy library to extract keywords and topics, and calculates sentiment scores using VADER Sentiment Analysis.

[1188] For video data, frame analysis and speech-to-text conversion are performed, for example, by using OpenCV to extract frame-by-frame features and Google Cloud Speech-to-Text to convert the speech to text.

[1189] 4. Training the predictive model and making predictions

[1190] The server trains a machine learning model based on the analyzed feature data, for example, a linear regression model from scikit-learn or a neural network from TensorFlow.

[1191] Use the trained model to predict the number of views for new content.

[1192] 5. Providing prediction results

[1193] The server generates prediction results in JSON format and sends them back to the user's device.

[1194] Terminal handling

[1195] The terminal provides an interface for users to access the system, input content, and check prediction results.

[1196] 1. Content input

[1197] The user uses the device to enter the title, text, tags, and metadata for the article or video.

[1198] 2. Send a prediction request

[1199] The terminal converts the input content data into JSON format and sends it to the server.

[1200] The prediction results returned from the server are received and displayed to the user.

[1201] User operations

[1202] The user accesses the system through a terminal and optimizes the content in the following steps.

[1203] 1. Content Creation

[1204] The user uses the device to create new article or video content. For example, the user might enter the following as the title: "Predicting the future with AI" and the text: "It is now possible to predict the future using AI technology. In particular, technology that can predict the number of content views with high accuracy is attracting attention."

[1205] 2. Prediction Request

[1206] The user clicks the button to request a prediction of the number of views of the content, and the terminal transmits the data to the server.

[1207] 3. Check and correct the prediction results

[1208] The server performs the analysis and predictions and sends the results back to the device. For example, results such as "Expected number of views: 1500 PV" and "Recommended title: 'How to predict the future with AI technology'" are displayed.

[1209] Users can then modify and optimize the title and body of their content based on the prediction results, and then publish the optimized content.

[1210] Through this system, users will be able to create effective content based on scientific evidence, helping to create new hit content.

[1211] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1212] Step 1: Data collection

[1213] Input: URLs of public data on the internet (news sites, blogs, social media posts, etc.)

[1214] How it works: The server sends an HTTP request to a URL to retrieve data in HTML or JSON format, using a web scraping tool (e.g., BeautifulSoup, Scrapy) or an API (e.g., Twitter API, YouTube Data API).

[1215] Output: Collected HTML data and JSON format data

[1216] Step 2: Data Preprocessing - Removing HTML Tags and Normalizing Special Characters

[1217] Input: Collected HTML data or JSON format data

[1218] What it does: The server uses BeautifulSoup to strip HTML tags and regular expressions to normalize special characters.

[1219] Output: Cleaned text data

[1220] Step 3: Data Preprocessing - Data Filtering

[1221] Input: Cleaned text data

[1222] What it does: The server removes data that does not contain certain keywords, for example, filtering out unwanted advertisements or spam posts.

[1223] Output: filtered text data

[1224] Step 4: Tokenization

[1225] Input: filtered text data

[1226] Specific operation: The server uses Python's NLTK library to split the text data into words.

[1227] Output: Tokenized text data

[1228] Step 5: Feature Extraction

[1229] Input: Tokenized text data

[1230] Specific operation: The server extracts features such as keywords, topics, and sentiment scores using natural language processing (NLP) techniques (e.g., spaCy, VADER Sentiment Analysis).

[1231] Output: Feature data such as keywords, topics, and sentiment scores

[1232] Step 6: Analyze the video data

[1233] Input: Video data

[1234] Specific operation: The server uses OpenCV to extract features for each frame and converts the audio into text using Google Cloud Speech-to-Text.

[1235] Output: Frame-by-frame feature data and speech-to-text data

[1236] Step 7: Train the machine learning model

[1237] Input: Feature data (keywords, topics, sentiment scores, etc.) and filtered data

[1238] Specific operation: The server trains a machine learning model based on the feature data using a linear regression model from scikit-learn or a neural network from TensorFlow.

[1239] Output: A trained machine learning model

[1240] Step 8: Predict content views

[1241] Input: New content feature data (keywords, topics, sentiment scores, etc.)

[1242] What it does: The server uses a trained machine learning model to predict the number of views for new content.

[1243] Output: Predicted number of views

[1244] Step 9: Providing prediction results

[1245] Input: Estimated number of views

[1246] Specific operation: The server formats the prediction results in JSON format and sends them to the device.

[1247] Output: Formatted prediction results (JSON format)

[1248] Step 10: Enter content

[1249] Input: Users enter content such as article or video titles, text, tags, and metadata.

[1250] What happens: The device takes these inputs and prepares the data for the prediction request.

[1251] Output: Input content data (JSON format)

[1252] Step 11: Send a prediction request

[1253] Input: Input content data (JSON format)

[1254] Specific operation: The terminal sends content data formatted in JSON format to the server via an HTTP POST request.

[1255] Output: Request sent

[1256] Step 12: Receive and display results

[1257] Input: Prediction results returned from the server (JSON format)

[1258] Specific behavior: The device receives the prediction results and displays them to the user. The device's UI components are used to visually display the results, such as the predicted number of views and recommended titles.

[1259] Output: Prediction results displayed to the user

[1260] This allows users to efficiently collect data, preprocess and analyze it, use machine learning models to accurately predict the number of views, and quickly provide the results to users.

[1261] (Application example 1)

[1262] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1263] While there are many systems that use publicly available data on the Internet to predict content view counts, most of them only provide prediction results and do not include specific optimization suggestions. As a result, users lack the information to effectively optimize titles and thumbnails based on the prediction results, and must rely on trial and error. This makes it difficult to maximize content views.

[1264] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1265] In this invention, the server includes means for collecting public data on the Internet, means for preprocessing and analyzing the collected data, means for predicting the number of views of content based on the preprocessed and analyzed data, means for optimizing the title and thumbnail of the content based on the predicted number of views, and means for providing the prediction results and optimization suggestions to users, thereby enabling users to efficiently and effectively optimize their content and maximize the number of views.

[1266] "Public data on the internet" refers to digital information that is freely accessible from networked sources, such as news sites, blogs, public API data, and social media posts.

[1267] "Methods of collection" refers to the techniques and methods used to obtain data from designated sources, including web scraping and the use of APIs.

[1268] "Preprocessing" refers to a series of operations performed to transform collected data into a format that is easier to analyze, such as text cleaning, normalizing special characters, and filtering out irrelevant data.

[1269] "Analytical means" refers to techniques and methods for analyzing collected and pre-processed data, including the application of natural language processing techniques and machine learning algorithms.

[1270] "Methods for predicting views" refers to techniques or methods for predicting the number of views of content based on analyzed data, including models such as regression analysis and neural networks.

[1271] "Optimization means" refers to technologies and methods for effectively changing or suggesting content titles and thumbnails based on the prediction results.

[1272] "Means of providing" refers to the interface and functions for notifying and displaying prediction results and optimization suggestions to users, including the user interface and notification functions.

[1273] This system effectively maximizes the number of content views by collecting, preprocessing, and analyzing public data on the Internet and providing users with prediction results and optimization suggestions. This system is mainly composed of three entities: a server, a terminal, and a user.

[1274] Server Processing

[1275] The server has the following functions:

[1276] 1. Data Collection:

[1277] The server uses web scraping tools and APIs to collect public data from the internet, such as news sites, blogs, and social media posts.

[1278] 2. Data Preprocessing:

[1279] The server performs text cleaning on the collected data, specifically removing HTML tags, normalizing special characters, and filtering out irrelevant data.

[1280] 3. Analysis and feature extraction:

[1281] The server uses natural language processing (NLP) techniques to extract features from the text data, including keywords, topics, and sentiment analysis results. It also uses the Python library Scikit-learn to vectorize the text data using TF-IDF.

[1282] 4. Views forecast:

[1283] The server trains a machine learning model based on the pre-processed and feature-extracted data to predict the number of views. It uses algorithms such as regression analysis and neural networks to build the predictive model.

[1284] 5. Optimization:

[1285] Based on the prediction results, the server generates suggestions to optimize the title and thumbnail of the content.

[1286] 6. Providing prediction results:

[1287] The server provides the prediction results and optimization suggestions to the user in real time.

[1288] Terminal handling

[1289] The terminal provides an interface for users to access the system and input content.

[1290] 1. Content Input:

[1291] Users use the device to input the title, text, tags, metadata, etc. of articles and videos.

[1292] 2. Send a prediction request:

[1293] The terminal converts the input content data into JSON format and sends it to the server.

[1294] 3. Receiving and displaying prediction results:

[1295] The terminal receives the prediction results and optimization suggestions returned from the server and displays them to the user.

[1296] User operations

[1297] The user accesses the system through a terminal and optimizes the content in the following steps.

[1298] 1. Content Creation:

[1299] Users use their devices to create new article or video content, for example by typing: "How to read the future with AI technology."

[1300] 2. Prediction request:

[1301] The user clicks the button requesting a prediction of the number of views of the content, and transmits data from the terminal to the server.

[1302] 3. Check and correct the predictions:

[1303] Users can check the prediction results and optimization suggestions displayed on their devices. For example, they may see "Expected page views: 1,500 PV" or "Recommended title: 'How to predict the future with AI technology.'" Based on this, users can revise the title and thumbnail and publish the optimized content.

[1304] Example prompt sentence:

[1305] The system collects and preprocesses data in the following way, then analyzes it to predict the number of views of the content. The specific steps are as follows:

[1306] 1. Collect news sites, blogs, and social media posts using web scraping tools and APIs.

[1307] 2. Clean the collected data and convert it into a suitable format for analysis.

[1308] 3. Extract features using natural language processing techniques.

[1309] 4. Train a machine learning model to predict viewership.

[1310] 5. Provide users with real-time prediction results.

[1311] Users can generate optimal titles and descriptions for various content to maximize views. For example, a video title like "How to Read the Future with AI Technology" is expected to attract many viewers according to the predictive model.

[1312] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1313] Step 1: Data collection

[1314] The server collects public data from news sites, blogs, social media posts, and other sources on the internet. To do this, it uses a web scraping tool or API. Specifically, it sends HTTP requests to specified URLs and parses the resulting HTML data to extract usable information. A list of URLs is given as input, and text data is obtained as output.

[1315] Step 2: Data Preprocessing

[1316] The server preprocesses the collected text data by removing HTML tags, normalizing special characters, and filtering irrelevant data. Raw text data is given as input, and clean text data is obtained as output. Regular expressions and text cleaning libraries are used in this process.

[1317] Step 3: Feature extraction

[1318] The server extracts features using the preprocessed text data. To do this, it uses natural language processing (NLP) techniques. Specifically, it uses the Python library Scikit-learn to perform TF-IDF vectorization. Clean text data is given as input, and vectorized feature data is obtained as output.

[1319] Step 4: Train the machine learning model

[1320] The server uses the feature data to train a machine learning model. This uses algorithms such as regression analysis and neural networks. Specifically, it uses the LinearRegression model from Scikit-learn. Vectorized feature data and label data (past view counts) are given as input, and the trained model is obtained as output.

[1321] Step 5: Predicting views

[1322] The server uses the trained machine learning model to predict the number of views for new content data provided by the user. The server receives feature data of the new content as input and obtains the predicted number of views as output.

[1323] Step 6: Optimization suggestions

[1324] The server generates optimal title and thumbnail suggestions based on the prediction results, taking into account past data on similar methods that have generated large numbers of views. The predicted number of views and content feature data are given as input, and the output is an optimized suggestion.

[1325] Step 7: Providing prediction results and optimization suggestions

[1326] The server provides the prediction results and optimization suggestions to the user in real time. This information is displayed to the user through the terminal. The prediction results and optimization suggestions are given as input and displayed to the user as output.

[1327] Step 8: Optimize your content for users

[1328] The user modifies the title and thumbnail of the content based on the provided prediction results and optimization suggestions, which effectively maximizes the number of views of the content. The prediction results and optimization suggestions are given as input, and the optimized content is obtained as output.

[1329] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1330] Overall system overview

[1331] This system collects, preprocesses, and analyzes public data on the Internet, and then predicts the number of views of content based on that data. Furthermore, it incorporates an emotion engine that recognizes user emotions, allowing for predictions that take emotion data into account. This system is primarily comprised of three components: a server, a terminal, and a user.

[1332] Server Processing

[1333] The server has the following functions:

[1334] 1. Data Collection

[1335] The server periodically collects public data from the internet using web scraping tools and APIs.

[1336] The server retrieves text and media data from news sites, blogs, social media, etc. and stores it in storage.

[1337] 2. Data Preprocessing

[1338] The server removes HTML tags from the collected data, normalizes special characters, and converts them into clean text data.

[1339] The server filters out irrelevant data and noise and converts it into a format suitable for analysis.

[1340] 3. Feature Extraction

[1341] The server uses natural language processing (NLP) technology to extract features such as keywords, topics, and sentiment analysis results from the text data.

[1342] The server also performs frame analysis and speech-to-text conversion on video data to extract similar features.

[1343] 4. Analysis by Emotion Engine

[1344] The server uses an emotion engine to recognize emotions from text data about users and content, for example classifying positive, negative, and neutral emotions.

[1345] The server incorporates the emotion data into the pre-processing and feature extraction data.

[1346] 5. Training the predictive model and making predictions

[1347] The server uses the pre-processed and feature-extracted data and the sentiment data to create a training dataset for machine learning.

[1348] The server trains a view prediction model using a regression analysis model or a neural network.

[1349] The server stores the trained model and makes it available for real-time predictions.

[1350] 6. Providing prediction results

[1351] The server generates prediction results and provides them to users in real time, including predicted views, virality predictions, suggestions for optimal presentation and composition, and sentiment-based recommendations.

[1352] Terminal handling

[1353] The terminal provides an interface for users to access the system and input content.

[1354] 1. Content input

[1355] The user enters the title, text, tags, metadata, etc. of the article or video.

[1356] 2. Send a prediction request

[1357] The terminal converts the input content data into JSON format and sends it to the server.

[1358] 3. Check the prediction results

[1359] The terminal receives the prediction results returned from the server and displays them to the user.

[1360] User operations

[1361] The user accesses the system through a terminal and optimizes the content in the following steps.

[1362] 1. Content Creation

[1363] The user uses the device to create new article or video content. For example, the user might enter the following as the title: "Predicting the future with AI" and the text: "It is now possible to predict the future using AI technology. In particular, technology that can predict the number of content views with high accuracy is attracting attention."

[1364] 2. Prediction Request

[1365] The user clicks the button to request a prediction of the number of views of the content, and transmits the data to the server.

[1366] The server performs the analysis and predictions and sends the results back to the device.

[1367] 3. Check and correct the prediction results

[1368] The user checks the prediction results displayed on their device, such as "Expected number of views: 1500 PV," "Recommended title: 'How to read the future with AI technology,'" and "Emotion score: High."

[1369] The user can then modify the title and body of the content based on the prediction results and publish the optimized content.

[1370] Specific examples

[1371] As a concrete example, we show a series of steps in which a user creates a blog post, a server is used to predict the number of views of the post, and optimization is performed taking into account emotional data.

[1372] 1. A user creates a blog post using a device. The title is "How to predict the future with AI technology" and the body is "It is possible to predict the future using AI technology..."

[1373] 2. The user clicks the "View Count Forecast" button, and the device sends the content data to the server.

[1374] 3. The server analyzes the data and extracts emotion data using an emotion engine.

[1375] 4. The server uses the trained prediction model to predict the number of views and generates a prediction result.

[1376] 5. The server sends the prediction results back to the device. For example, "Expected number of views: 1500 PV", "Recommended title: 'How to predict the future with AI'", "Sentiment score: High".

[1377] 6. The user checks the prediction results on their device, changes the title to the recommended title, and publishes the optimized article.

[1378] Through the system of the present invention, users can create effective content based on scientific evidence that takes into account emotional data, thereby supporting the creation of new hit content.

[1379] The processing flow will be explained below.

[1380] Step 1: Data collection

[1381] The server periodically collects public data from the internet using web scraping tools and APIs.

[1382] The server retrieves text and media data from news sites, blogs, social media, etc. and stores it in storage.

[1383] Step 2: Data Preprocessing

[1384] The server removes HTML tags from the collected data, normalizes special characters, and converts them into clean text data.

[1385] The server filters out irrelevant data and noise and converts it into a format suitable for analysis.

[1386] Step 3: Feature extraction

[1387] The server uses natural language processing (NLP) technology to extract features such as keywords, topics, and sentiment analysis results from the text data.

[1388] The server also performs frame analysis and speech-to-text conversion on video data to extract similar features.

[1389] Step 4: Analysis by Emotion Engine

[1390] The server uses an emotion engine to recognize emotions from text data about users and content.

[1391] The server classifies and scores positive, negative, and neutral sentiment within the text data.

[1392] Step 5: Train the predictive model

[1393] The server uses the pre-processed and feature-extracted data and the sentiment data to create a training dataset for machine learning.

[1394] The server trains a view prediction model using a regression analysis model or a neural network.

[1395] The server stores the trained model and makes it available for real-time predictions.

[1396] Step 6: Enter content

[1397] The user uses the device interface to input the title, text, tags, metadata, etc. of the article or video.

[1398] Step 7: Send a prediction request

[1399] When the user clicks the "View Count Prediction" button, the device converts the content data into JSON format and sends it to the server.

[1400] Step 8: Data analysis and prediction

[1401] The server analyzes the received content data and performs pre-processing and feature extraction.

[1402] The server analyzes the emotion of the content using an emotion engine and calculates an emotion score.

[1403] The server uses a trained prediction model to predict the number of views of the content.

[1404] Step 9: Providing prediction results

[1405] The server generates and returns prediction results to the user, including expected views, virality predictions, suggestions for optimal wording and composition, and sentiment scores.

[1406] Step 10: Check and correct the prediction results

[1407] The user checks the prediction results received on their device, which may include, for example, "Expected number of views: 1500 PV," "Recommended title: 'How to read the future with AI technology,'" and "Emotion score: High."

[1408] The user can then modify the title and body of the content based on the prediction results.

[1409] The user publishes the optimized content.

[1410] Example 2

[1411] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1412] Conventional systems that use public data on the Internet to predict the number of views of content are limited to preprocessing and analysis of collected data, and it is difficult to make predictions that take user emotions into account. As a result, prediction accuracy is low and it is difficult to create effective content.

[1413] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for collecting public data on the Internet, means for preprocessing and analyzing the collected data, means for predicting the number of views of content based on the preprocessed and analyzed data, means for extracting and analyzing emotion data using an emotion engine that recognizes the user's emotions, and means for providing the prediction result to the user. This makes it possible to predict the number of views of content with high accuracy, taking the user's emotions into consideration.

[1414] "Public data on the Internet" refers to all information, including text data and media data, that is publicly available on the Internet.

[1415] "Means of collection" refers to the function of regularly obtaining publicly available data on the Internet using web scraping tools and APIs.

[1416] "Preprocessing" refers to the process of removing HTML tags and normalizing special characters from collected data to convert it into clean text data.

[1417] "Means of analysis" refers to the function of analyzing preprocessed data using natural language processing technology, etc., and extracting necessary features and information.

[1418] "Means for predicting the number of content views" refers to a function that uses a machine learning model to predict the number of content views based on pre-processed and analyzed data.

[1419] An "emotion engine" is software or a service for analyzing user emotions, and has the ability to recognize positive, negative, and neutral emotions from text data.

[1420] "Emotion data" is the result data of emotions related to users and content extracted by the emotion engine.

[1421] The "means for providing predicted results" refers to the function of the server to generate predicted results and provide them to the user in real time.

[1422] "Natural language processing technology" is a technology for analyzing text data and extracting features, and includes keyword extraction, topic modeling, sentiment analysis, etc.

[1423] "Content data entered by the user" refers to information such as the title, text, tags, and metadata of articles and videos that the user sends to the system via their terminal.

[1424] "Server" refers to a computer system that centrally handles functions such as data collection, preprocessing, analysis, emotion recognition, and view count prediction.

[1425] "Terminal" refers to the device or application that a User uses to access the System and input Content.

[1426] This system collects, preprocesses, and analyzes public data on the Internet, and predicts the number of views of content based on that data. Furthermore, it combines an emotion engine that recognizes user emotions, allowing for predictions that take emotion data into account.

[1427] Server functions and processing contents

[1428] 1. Data Collection

[1429] The server uses web scraping tools (e.g., BeautifulSoup, Scrapy) and APIs to collect public data from news sites, blogs, social media, etc. For example, article text from news sites and social media posts are stored in a database. The server periodically collects this data and stores it in storage.

[1430] 2. Data Preprocessing

[1431] The server converts the collected data into clean text data by removing HTML tags and normalizing special characters. Specifically, it uses BeautifulSoup to remove unnecessary tags and format the data in a unified format. It also filters out advertisements and irrelevant parts to create a dataset suitable for analysis.

[1432] 3. Feature Extraction

[1433] The server uses natural language processing (NLP) techniques (e.g., spaCy, NLTK) to extract features such as keywords, topics, and sentiment analysis results from text data. For example, it performs topic modeling and sentiment scoring. For video data, it performs frame analysis using OpenCV and speech-to-text conversion using Google Speech-to-Text to extract similar features.

[1434] 4. Analysis by Emotion Engine

[1435] The server analyzes the user's emotions using an emotion engine (e.g., Natural Language Understanding), which classifies positive, negative, and neutral emotions from the text data and integrates the emotion data into a feature dataset.

[1436] 5. Training the predictive model and making predictions

[1437] The server uses the preprocessed and feature-extracted dataset to train a machine learning model (e.g., TensorFlow, PyTorch). The trained predictive model is used to predict the number of views of new content in real time.

[1438] 6. Providing prediction results

[1439] The server uses the trained model to generate predictions and provide them to the user, such as generating JSON data containing predicted views and sentiment-based recommendations, and sending it back to the device.

[1440] User and Device Capabilities

[1441] 1. Content input

[1442] The user uses the device to enter the title, text, tags, and metadata of a new article or video. For example, the title might be "Predicting the future with AI" and the text might be "AI technology makes it possible to predict the future...".

[1443] 2. Send a prediction request

[1444] When the user clicks the "View Count Prediction" button, the device converts the entered content data into JSON format and sends it to the server.

[1445] 3. Check the prediction results

[1446] The device receives the prediction results sent from the server and displays them to the user, such as "Expected number of views: 1500 PV," a recommended title "How to predict the future with AI technology," and an emotion score of "High."

[1447] 4. Modify and optimize your content

[1448] The user checks the prediction results and modifies the content as necessary, for example, changing the title to a recommended title or adjusting the wording of the body text. The optimized content is then published by the user.

[1449] Examples and prompts

[1450] As a concrete example, we show a series of steps in which a user creates a blog post, a server is used to predict the number of views of the post, and optimization is performed taking into account emotional data.

[1451] Title: "How to predict the future with AI technology"

[1452] Body text: "AI technology is making it possible to predict the future. In particular, technology that can predict the number of content views with high accuracy is attracting attention."

[1453] An example prompt might be, "What is the expected number of views for this article?"

[1454] Through this system, users will be able to create effective content based on scientific evidence, helping to create new hit content.

[1455] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1456] Step 1: Data collection

[1457] Input: The server retrieves public data from the internet (e.g., news articles, blog posts, social media posts, etc.) from a specified URL or API.

[1458] What it does: The server collects public data using web scraping tools (e.g., BeautifulSoup, Scrapy) or APIs. For example, it scrapes articles from news sites with BeautifulSoup and retrieves social media posts through APIs.

[1459] Output: The collected raw data is stored in the server's storage. It includes files in various formats, such as text data, image data, and audio data.

[1460] Step 2: Data Preprocessing

[1461] Input: Raw data collected by the server.

[1462] What it does: The server uses BeautifulSoup to strip HTML tags, normalize special characters, and perform noise filtering to remove irrelevant data, such as removing ads and script tags from HTML and resolving encoding issues.

[1463] Output: Clean text data and structured data. For example, ads and irrelevant information are removed and a normalized text file is output.

[1464] Step 3: Feature extraction

[1465] Input: Clean text data.

[1466] Specific operation: The server uses natural language processing (NLP) techniques (e.g., spaCy, NLTK) to extract features such as keywords, topics, and sentiment analysis results. For video data, it performs frame analysis using OpenCV and converts the audio to text using Google Speech-to-Text.

[1467] Output: A dataset of extracted features, including a keyword list and sentiment scores from text data, and important scenes extracted from video frames.

[1468] Step 4: Analysis by Emotion Engine

[1469] Input: Extracted feature dataset.

[1470] Specific operation: The server uses an emotion engine (e.g., Natural Language Understanding) to classify the emotion of each piece of text data and calculate an emotion score, for example, classifying it as "positive," "negative," or "neutral."

[1471] Output: A feature dataset with sentiment data added, e.g., sentiment scores for each text data.

[1472] Step 5: Train the predictive model and make predictions

[1473] Input: Preprocessed and feature extracted dataset.

[1474] What it does: The server uses machine learning models (e.g., TensorFlow, PyTorch) to train regression analysis or neural network models using this data, and then uses the trained models to predict the number of views for new content.

[1475] Output: The trained prediction model and prediction results, such as predicted views and suggested titles for new articles.

[1476] Step 6: Providing prediction results

[1477] Input: JSON data containing prediction results.

[1478] Specific operation: The server generates a prediction result and sends it back to the device. For example, "Expected number of views: 1500 PV" or "Recommended title: 'How to predict the future with AI technology'".

[1479] Output: The prediction result is sent to the device so that the user can check it.

[1480] Step 7: User input and prediction request

[1481] Input: User-generated content data (article title, body, tags, metadata, etc.).

[1482] Specific operation: The user inputs content through the terminal and clicks the "View Count Prediction" button. The terminal converts the content data into JSON format and sends it to the server.

[1483] Output: The content data sent to the server.

[1484] Step 8: Check the prediction results and modify the content

[1485] Input: Prediction results sent from the server.

[1486] Specific operation: The user checks the prediction results on the device and modifies the content as necessary. For example, the title can be changed to "How to read the future with AI technology" based on the prediction results.

[1487] Output: The optimized content is published.

[1488] As described above, data input, data processing, data calculation, and output are performed in each processing step, enabling users to create effective content by utilizing highly accurate prediction results and emotion data.

[1489] (Application example 2)

[1490] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1491] Conventional content view count prediction systems simply predicted the number of views without considering user emotions, making it difficult to optimize content more effectively. In particular, in the advertising field, where emotion-based optimization is required, there was a lack of functionality to suggest optimal titles and content.

[1492] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1493] In this invention, the server includes means for collecting public data on the Internet, means for preprocessing and analyzing the collected data, means for predicting the number of views of content based on the preprocessed and analyzed data, means for recognizing user emotions using a sentiment analysis engine and making predictions taking the emotional data into consideration, means for providing the prediction results and optimization suggestions to the user, and means for generating optimal titles and content revision suggestions for advertising content, thereby enabling the prediction and optimization of the number of views of advertising content that reflects user emotions.

[1494] "Public data on the internet" refers to information that is freely accessible on websites, blogs, social media platforms, etc.

[1495] "Means of collection" refers to the process of using web scraping tools or APIs to obtain public data on a server.

[1496] "Preprocessing" refers to the process of removing unnecessary HTML tags from collected data, normalizing special characters, and converting the data into a format suitable for analysis.

[1497] "Means of analysis" refers to the process of analyzing data using natural language processing techniques and extracting features such as keywords, topics, and sentiment.

[1498] "Method for predicting content views" refers to the process of using a machine learning model to predict content views based on pre-processed and analyzed data.

[1499] "Sentiment analysis engine" refers to an algorithm that recognizes positive, negative, and neutral emotions from text data entered by the user.

[1500] "Means for making predictions taking into account emotional data" refers to a process of incorporating emotional data into a view count prediction model to make view count predictions that take into account the influence of emotions.

[1501] "Means for providing prediction results and optimization suggestions to users" refers to a process for providing users with prediction results as well as suggestions for more effective content expression and configuration.

[1502] "Means for generating optimal titles and suggested content modifications for advertising content" refers to the process of generating effective titles and suggested text modifications for advertising content based on prediction results and emotional data.

[1503] The present invention relates to a system that collects, preprocesses, and analyzes public data on the Internet and predicts the number of views of content based on that data. This system, combined with a sentiment analysis engine, performs advanced predictions that take emotional data into account and provides optimal titles and content revisions, particularly in the advertising field. A specific embodiment of this system is described below.

[1504] Server Processing

[1505] The server has the following main functions:

[1506] 1. Data Collection

[1507] The server periodically collects public data from the internet using web scraping tools and APIs.

[1508] For example, text and media data is acquired from news sites, blogs, and social media and stored in a database.

[1509] 2. Data Preprocessing

[1510] The server removes HTML tags from the collected data, normalizes special characters, and converts them into clean text data.

[1511] It also filters out irrelevant data and noise and converts it into a format suitable for analysis.

[1512] 3. Feature Extraction

[1513] The server uses natural language processing technology (e.g., spaCy or NLTK) to extract features such as keywords, topics, and sentiment analysis results from the text data.

[1514] In the case of video data, frame analysis and speech-to-text conversion are performed to extract similar features.

[1515] 4. Analysis using a sentiment analysis engine

[1516] The server uses a sentiment analysis engine (such as VADER) to classify the user's sentiment from the text data.

[1517] The classified sentiment data (positive, negative, neutral) is incorporated into the preprocessing and feature extraction data.

[1518] 5. Training the predictive model and making predictions

[1519] The server uses the preprocessed and feature-extracted data and sentiment data to train a machine learning model (e.g., Scikit-learn or TensorFlow).

[1520] The trained model is used to predict the number of views of the content.

[1521] 6. Providing prediction results and optimization suggestions

[1522] The server generates and provides the user with optimal representation and composition suggestions along with the prediction results.

[1523] Specifically, it generates optimal titles and suggested content revisions for advertising content, reflecting user emotional data.

[1524] Terminal handling

[1525] The terminal provides an interface for users to access the system and input content.

[1526] 1. Content input

[1527] Users enter the title, text, tags, metadata, etc. of the article or video.

[1528] 2. Send a prediction request

[1529] The terminal converts the input content data into JSON format and sends it to the server.

[1530] 3. Check the prediction results and optimization suggestions

[1531] The terminal receives the prediction results and optimization suggestions returned from the server and displays them to the user.

[1532] User operations

[1533] A user accesses the system through a terminal and optimizes advertising content by following the steps below.

[1534] 1. Content Creation

[1535] The user uses the device to create new advertising content, for example, by typing "We've launched a new smartwatch. Learn more!"

[1536] 2. Prediction Request

[1537] The user clicks the button to request a prediction of the number of views of the content, and transmits the data to the server.

[1538] 3. Check and modify the prediction results and optimization suggestions

[1539] Users can view the prediction results and optimization suggestions displayed on their device, such as "Expected page views: 3,000 PV," "Recommended title: 'Limited sale! Get it now,'" and "Sentiment score: High."

[1540] Users optimize and publish content based on prediction results and suggestions.

[1541] Specific examples

[1542] When a user submits an ad to the system saying, "We've released a new smartwatch. Click here for details!", the system performs sentiment analysis and predicts the number of views. The predicted results are displayed as "Sentiment score: Medium," "Expected number of views: 2000 PV," and "Recommended title: 'Limited time only! Check out the details of the limited edition smartwatch.'" The user can then change the displayed title to optimize the ad.

[1543] Prompt Sentence Examples

[1544] Prompt statement:

[1545] The ad data entered by the user is listed below. Please perform sentiment analysis and respond with a view prediction and optimal ad modification suggestions.

[1546] Ad title: New smartwatch released

[1547] Ad text: We've launched a new smartwatch. Learn more here!

[1548] Tags: smartwatch sale

[1549] Expected output:

[1550] Estimated number of views: XXXX PV

[1551] Recommended title: 'Now only! Check out the details of the limited edition smartwatch'

[1552] Sentiment score: High / Medium / Low

[1553] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1554] Step 1:

[1555] Data collection (server)

[1556] Input: Public data from the internet (e.g., news articles, blog posts, social media posts).

[1557] What it does: The server periodically collects this public data using web scraping tools and APIs and stores it in a database.

[1558] Output: The raw data collected.

[1559] Step 2:

[1560] Data preprocessing (server)

[1561] Input: The raw data collected.

[1562] What it does: The server strips HTML tags, normalizes special characters, filters irrelevant data and noise, and generates clean text data.

[1563] Output: Preprocessed and clean text data.

[1564] Step 3:

[1565] Feature extraction (server)

[1566] Input: Preprocessed clean text data.

[1567] Specific operation: The server uses natural language processing techniques (e.g., spaCy or NLTK) to extract features such as keywords, topics, and sentiment analysis results from text data. For video data, it also performs frame analysis and speech-to-text conversion.

[1568] Output: Feature-extracted data (keywords, topics, sentiment scores, etc.).

[1569] Step 4:

[1570] Sentiment analysis (server)

[1571] Input: Feature extracted data.

[1572] What happens: The server uses a sentiment analysis engine (e.g., VADER) to classify specific sentiments (positive, negative, neutral) and calculate a sentiment score.

[1573] Output: Data with sentiment scores.

[1574] Step 5:

[1575] Training and predicting the predictive model (server)

[1576] Input: Preprocessed and sentiment-scored data.

[1577] What it does: The server uses a machine learning model (e.g., Scikit-learn or TensorFlow) to train a model to predict the number of content views and makes the prediction.

[1578] Output: View count prediction results.

[1579] Step 6:

[1580] Generation of prediction results and optimization proposals (server)

[1581] Input: View count prediction results and sentiment score.

[1582] Specific operation: Based on the prediction results, the server generates optimal ad titles and content revision suggestions and provides them to the user.

[1583] Output: Forecast results with optimization suggestions.

[1584] Step 7:

[1585] Content input (terminal)

[1586] Input: Ad content (e.g. title, body, tags).

[1587] Specific actions: The user uses the device to enter data into the content input interface and clicks the submit button.

[1588] Output: The input content data is sent to the server.

[1589] Step 8:

[1590] Sending a prediction request (device)

[1591] Input: The content data entered.

[1592] Specific operation: The terminal converts the content data into JSON format and sends it to the server.

[1593] Output: The content data sent to the server.

[1594] Step 9:

[1595] Check prediction results and optimization suggestions (device)

[1596] Input: Prediction results and optimization suggestions returned by the server.

[1597] Specific operation: The device receives the prediction results and optimization suggestions returned from the server and displays them to the user.

[1598] Output: Prediction results and optimization suggestions displayed to the user.

[1599] Step 10:

[1600] Content optimization and publishing (user)

[1601] Input: Prediction results and optimization suggestions.

[1602] What it does: The user modifies the content based on the displayed predictions and suggestions, then clicks the publish button to publish the optimized content.

[1603] Output: The published optimized content.

[1604] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1605] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1606] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1607] [Fourth embodiment]

[1608] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1609] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1610] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1611] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1612] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1613] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1614] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1615] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1616] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1617] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1618] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1619] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1620] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1621] Overall system overview

[1622] This invention is a system that collects, preprocesses, and analyzes public data on the Internet and predicts the number of views of content based on the collected data. The system is mainly composed of three entities: a server, a terminal, and a user.

[1623] Server Processing

[1624] The server has the following functions:

[1625] 1. Data Collection

[1626] The server uses web scraping tools and APIs to collect public data from the internet.

[1627] Collections include news sites, blogs, social media posts, and more.

[1628] 2. Data Preprocessing

[1629] The server performs text cleaning on the collected data, specifically removing HTML tags, normalizing special characters, and filtering out irrelevant data.

[1630] The cleaned data is then tokenized and converted into a format suitable for analysis.

[1631] 3. Analysis and feature extraction

[1632] The server uses natural language processing (NLP) techniques to extract features from the text data, including keywords, topics, and sentiment analysis results.

[1633] The server also performs frame analysis and speech-to-text conversion on video data to extract similar features.

[1634] 4. Training the predictive model and making predictions

[1635] The server uses this data to train machine learning models, which use algorithms such as regression analysis and neural networks.

[1636] The trained model is used to predict the number of views for different content.

[1637] 5. Providing prediction results

[1638] The server generates prediction results and provides them to users in real time, including predicted views, diffusion forecasts, and suggestions for optimal presentation and composition.

[1639] Terminal handling

[1640] The terminal provides an interface for users to access the system and input content.

[1641] 1. Content input

[1642] The user enters the title, text, tags, metadata, etc. of the article or video.

[1643] 2. Send a prediction request

[1644] The terminal converts the input content data into JSON format and sends it to the server.

[1645] The prediction results returned from the server are received and displayed to the user.

[1646] User operations

[1647] The user accesses the system through a terminal and optimizes the content in the following steps.

[1648] 1. Content Creation

[1649] The user uses the device to create new article or video content. For example, the user might enter the following as the title: "Predicting the future with AI" and the text: "It is now possible to predict the future using AI technology. In particular, technology that can predict the number of content views with high accuracy is attracting attention."

[1650] 2. Prediction Request

[1651] The user clicks the button to request a prediction of the number of views of the content, and transmits the data to the server.

[1652] The server performs the analysis and predictions and sends the results back to the device.

[1653] 3. Check and correct the prediction results

[1654] The user can check the prediction results displayed on their device, such as "Expected number of views: 1500 PV" or "Recommended title: 'How to read the future with AI technology'."

[1655] The user can then modify the title and body of the content based on the prediction results and publish the optimized content.

[1656] Specific examples

[1657] As a concrete example, we will show a series of steps in which a user creates a blog article, and then uses a server to predict and optimize the number of views of that article.

[1658] 1. A user creates a blog post using a device. The title is "How to predict the future with AI technology" and the body is "It is possible to predict the future using AI technology..."

[1659] 2. The user clicks the "View Count Forecast" button, and the device sends the content data to the server.

[1660] 3. The server analyzes the data and uses a predictive model to predict the number of views.

[1661] 4. The server generates a prediction result and sends it back to the device. For example, "Expected number of views: 1500 PV" and "Recommended title: 'How to predict the future with AI'."

[1662] 5. The user checks the prediction results on their device, changes the title to the recommended title, and publishes the article.

[1663] Through the system of the present invention, users can create effective content based on scientific evidence, which can help create new hit content.

[1664] The processing flow will be explained below.

[1665] Step 1: Data collection

[1666] The server periodically collects public data from the internet using web scraping tools and APIs.

[1667] The server retrieves text and media data from news sites, blogs, social media, etc. and stores it in storage.

[1668] Step 2: Data Preprocessing

[1669] The server removes HTML tags from the collected data, normalizes special characters, and converts them into clean text data.

[1670] The server filters out irrelevant data and noise and converts it into a format suitable for analysis.

[1671] Step 3: Feature extraction

[1672] The server uses natural language processing (NLP) technology to extract features such as keywords, topics, and sentiment analysis results from the text data.

[1673] The server also performs frame analysis and speech-to-text conversion on video data to extract similar features.

[1674] Step 4: Training the predictive model

[1675] The server uses the pre-processed and feature-extracted data to create a training dataset for machine learning.

[1676] The server trains a view prediction model using a regression analysis model or a neural network.

[1677] The server stores the trained model and makes it available for real-time predictions.

[1678] Step 5: Enter content

[1679] The user uses the device interface to input the title, text, tags, metadata, etc. of the article or video.

[1680] Step 6: Send a prediction request

[1681] When the user clicks the "View Count Prediction" button, the device converts the content data into JSON format and sends it to the server.

[1682] Step 7: Data analysis and prediction

[1683] The server analyzes the received content data and performs pre-processing and feature extraction.

[1684] The server uses a trained prediction model to predict the number of views of the content.

[1685] Step 8: Providing prediction results

[1686] The server generates prediction results and returns them to the user, including predicted views, diffusion forecasts, and suggestions for optimal presentation and composition.

[1687] Step 9: Check and correct the predictions

[1688] The user checks the prediction results received on their device, such as "Expected number of views: 1500 PV" and "Recommended title: 'How to read the future with AI technology.'"

[1689] The user can then modify the title and body of the content based on the prediction results and publish the optimized content.

[1690] Example 1

[1691] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1692] In conventional content view prediction systems, the processes from data collection to preprocessing, analysis, prediction, and result provision are often not performed consistently and efficiently. Also, bias toward specific analysis methods or models can make accurate predictions difficult. Another issue is the lack of a means to efficiently transmit content data entered by users to a server in real time and quickly return prediction results. There is a need for a more accurate and efficient content view prediction system that can resolve these issues.

[1693] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1694] In this invention, the server includes means for collecting public data on the Internet, means for preprocessing the collected data, including removing HTML tags, normalizing special characters, and filtering irrelevant data, means for tokenizing the preprocessed data and converting it into a format suitable for analysis, means for extracting features such as keywords, topics, and sentiment analysis results from the text data using natural language processing techniques, means for training a machine learning model based on the analyzed data and predicting the number of views of content, and means for providing the predicted results to users in real time. This makes it possible to efficiently collect data, perform preprocessing and analysis, accurately predict the number of views using the machine learning model, and quickly provide the results to users.

[1695] A "server" is a computer system that collects, analyzes, and predicts public data on a network and provides the results to users.

[1696] "Public data on the internet" is information that is freely accessible online, such as news sites, blogs, and social media posts.

[1697] "Preprocessing" refers to data cleaning tasks such as removing HTML tags from collected data, normalizing special characters, and filtering out irrelevant data.

[1698] "Tokenization" is the process of breaking down text data into units of words and phrases and converting them into a format suitable for analysis.

[1699] "Natural language processing technology" is a technology for extracting features such as keywords, topics, and emotions from text data.

[1700] A "machine learning model" is an algorithm or program that is trained on data to perform a specific task (in this case, predicting content views).

[1701] "Real-time" refers to a form of immediate response where the time between when a user sends data and when they receive the results is very short.

[1702] A "terminal" is a computer or device that allows a user to access the system, enter content, and view results.

[1703] "Content" is a collection of data including article or video titles, text, tags, metadata, etc.

[1704] MODE FOR CARRYING OUT THE INVENTION

[1705] This system collects public data on the Internet, preprocesses and analyzes it, and then predicts the number of views of content based on the collected data. The system is mainly composed of three entities: a server, a terminal, and a user.

[1706] Server Processing

[1707] The server performs data processing and data calculation using the following hardware and software.

[1708] 1. Data Collection

[1709] The server uses web scraping tools (e.g., BeautifulSoup, Scrapy) and APIs (e.g., Twitter API, YouTube Data API) to collect data from news sites, blogs, social media posts, etc.

[1710] Specifically, the server sends an HTTP request to the specified URL and retrieves HTML data or JSON format data.

[1711] 2. Data Preprocessing

[1712] The server performs text cleaning such as removing HTML tags, normalizing special characters, and filtering out irrelevant data. For example, it uses BeautifulSoup to remove HTML tags and regular expressions to normalize special characters.

[1713] The cleaned data is then tokenized and converted into a format suitable for analysis, for example by splitting the text into words using Python's NLTK library.

[1714] 3. Analysis and feature extraction

[1715] The server analyzes the text data using natural language processing (NLP) techniques to extract features such as keywords, topics, and sentiment analysis results. For example, it uses the spaCy library to extract keywords and topics, and calculates sentiment scores using VADER Sentiment Analysis.

[1716] For video data, frame analysis and speech-to-text conversion are performed, for example, by using OpenCV to extract frame-by-frame features and Google Cloud Speech-to-Text to convert the speech to text.

[1717] 4. Training the predictive model and making predictions

[1718] The server trains a machine learning model based on the analyzed feature data, for example, a linear regression model from scikit-learn or a neural network from TensorFlow.

[1719] Use the trained model to predict the number of views for new content.

[1720] 5. Providing prediction results

[1721] The server generates prediction results in JSON format and sends them back to the user's device.

[1722] Terminal handling

[1723] The terminal provides an interface for users to access the system, input content, and check prediction results.

[1724] 1. Content input

[1725] The user uses the device to enter the title, text, tags, and metadata for the article or video.

[1726] 2. Send a prediction request

[1727] The terminal converts the input content data into JSON format and sends it to the server.

[1728] The prediction results returned from the server are received and displayed to the user.

[1729] User operations

[1730] The user accesses the system through a terminal and optimizes the content in the following steps.

[1731] 1. Content Creation

[1732] The user uses the device to create new article or video content. For example, the user might enter the following as the title: "Predicting the future with AI" and the text: "It is now possible to predict the future using AI technology. In particular, technology that can predict the number of content views with high accuracy is attracting attention."

[1733] 2. Prediction Request

[1734] The user clicks the button to request a prediction of the number of views of the content, and the terminal transmits the data to the server.

[1735] 3. Check and correct the prediction results

[1736] The server performs the analysis and predictions and sends the results back to the device. For example, results such as "Expected number of views: 1500 PV" and "Recommended title: 'How to predict the future with AI technology'" are displayed.

[1737] Users can then modify and optimize the title and body of their content based on the prediction results, and then publish the optimized content.

[1738] Through this system, users will be able to create effective content based on scientific evidence, helping to create new hit content.

[1739] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1740] Step 1: Data collection

[1741] Input: URLs of public data on the internet (news sites, blogs, social media posts, etc.)

[1742] How it works: The server sends an HTTP request to a URL to retrieve data in HTML or JSON format, using a web scraping tool (e.g., BeautifulSoup, Scrapy) or an API (e.g., Twitter API, YouTube Data API).

[1743] Output: Collected HTML data and JSON format data

[1744] Step 2: Data Preprocessing - Removing HTML Tags and Normalizing Special Characters

[1745] Input: Collected HTML data or JSON format data

[1746] What it does: The server uses BeautifulSoup to strip HTML tags and regular expressions to normalize special characters.

[1747] Output: Cleaned text data

[1748] Step 3: Data Preprocessing - Data Filtering

[1749] Input: Cleaned text data

[1750] What it does: The server removes data that does not contain certain keywords, for example, filtering out unwanted advertisements or spam posts.

[1751] Output: filtered text data

[1752] Step 4: Tokenization

[1753] Input: filtered text data

[1754] Specific operation: The server uses Python's NLTK library to split the text data into words.

[1755] Output: Tokenized text data

[1756] Step 5: Feature Extraction

[1757] Input: Tokenized text data

[1758] Specific operation: The server extracts features such as keywords, topics, and sentiment scores using natural language processing (NLP) techniques (e.g., spaCy, VADER Sentiment Analysis).

[1759] Output: Feature data such as keywords, topics, and sentiment scores

[1760] Step 6: Analyze the video data

[1761] Input: Video data

[1762] Specific operation: The server uses OpenCV to extract features for each frame and converts the audio into text using Google Cloud Speech-to-Text.

[1763] Output: Frame-by-frame feature data and speech-to-text data

[1764] Step 7: Train the machine learning model

[1765] Input: Feature data (keywords, topics, sentiment scores, etc.) and filtered data

[1766] Specific operation: The server trains a machine learning model based on the feature data using a linear regression model from scikit-learn or a neural network from TensorFlow.

[1767] Output: A trained machine learning model

[1768] Step 8: Predict content views

[1769] Input: New content feature data (keywords, topics, sentiment scores, etc.)

[1770] What it does: The server uses a trained machine learning model to predict the number of views for new content.

[1771] Output: Predicted number of views

[1772] Step 9: Providing prediction results

[1773] Input: Estimated number of views

[1774] Specific operation: The server formats the prediction results in JSON format and sends them to the device.

[1775] Output: Formatted prediction results (JSON format)

[1776] Step 10: Enter content

[1777] Input: Users enter content such as article or video titles, text, tags, and metadata.

[1778] What happens: The device takes these inputs and prepares the data for the prediction request.

[1779] Output: Input content data (JSON format)

[1780] Step 11: Send a prediction request

[1781] Input: Input content data (JSON format)

[1782] Specific operation: The terminal sends content data formatted in JSON format to the server via an HTTP POST request.

[1783] Output: Request sent

[1784] Step 12: Receive and display results

[1785] Input: Prediction results returned from the server (JSON format)

[1786] Specific behavior: The device receives the prediction results and displays them to the user. The device's UI components are used to visually display the results, such as the predicted number of views and recommended titles.

[1787] Output: Prediction results displayed to the user

[1788] This allows users to efficiently collect data, preprocess and analyze it, use machine learning models to accurately predict the number of views, and quickly provide the results to users.

[1789] (Application example 1)

[1790] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1791] While there are many systems that use publicly available data on the Internet to predict content view counts, most of them only provide prediction results and do not include specific optimization suggestions. As a result, users lack the information to effectively optimize titles and thumbnails based on the prediction results, and must rely on trial and error. This makes it difficult to maximize content views.

[1792] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1793] In this invention, the server includes means for collecting public data on the Internet, means for preprocessing and analyzing the collected data, means for predicting the number of views of content based on the preprocessed and analyzed data, means for optimizing the title and thumbnail of the content based on the predicted number of views, and means for providing the prediction results and optimization suggestions to users, thereby enabling users to efficiently and effectively optimize their content and maximize the number of views.

[1794] "Public data on the internet" refers to digital information that is freely accessible from networked sources, such as news sites, blogs, public API data, and social media posts.

[1795] "Methods of collection" refers to the techniques and methods used to obtain data from designated sources, including web scraping and the use of APIs.

[1796] "Preprocessing" refers to a series of operations performed to transform collected data into a format that is easier to analyze, such as text cleaning, normalizing special characters, and filtering out irrelevant data.

[1797] "Analytical means" refers to techniques and methods for analyzing collected and pre-processed data, including the application of natural language processing techniques and machine learning algorithms.

[1798] "Methods for predicting views" refers to techniques or methods for predicting the number of views of content based on analyzed data, including models such as regression analysis and neural networks.

[1799] "Optimization means" refers to technologies and methods for effectively changing or suggesting content titles and thumbnails based on the prediction results.

[1800] "Means of providing" refers to the interface and functions for notifying and displaying prediction results and optimization suggestions to users, including the user interface and notification functions.

[1801] This system effectively maximizes the number of content views by collecting, preprocessing, and analyzing public data on the Internet and providing users with prediction results and optimization suggestions. This system is mainly composed of three entities: a server, a terminal, and a user.

[1802] Server Processing

[1803] The server has the following functions:

[1804] 1. Data Collection:

[1805] The server uses web scraping tools and APIs to collect public data from the internet, such as news sites, blogs, and social media posts.

[1806] 2. Data Preprocessing:

[1807] The server performs text cleaning on the collected data, specifically removing HTML tags, normalizing special characters, and filtering out irrelevant data.

[1808] 3. Analysis and feature extraction:

[1809] The server uses natural language processing (NLP) techniques to extract features from the text data, including keywords, topics, and sentiment analysis results. It also uses the Python library Scikit-learn to vectorize the text data using TF-IDF.

[1810] 4. Views forecast:

[1811] The server trains a machine learning model based on the pre-processed and feature-extracted data to predict the number of views. It uses algorithms such as regression analysis and neural networks to build the predictive model.

[1812] 5. Optimization:

[1813] Based on the prediction results, the server generates suggestions to optimize the title and thumbnail of the content.

[1814] 6. Providing prediction results:

[1815] The server provides the prediction results and optimization suggestions to the user in real time.

[1816] Terminal handling

[1817] The terminal provides an interface for users to access the system and input content.

[1818] 1. Content Input:

[1819] Users use the device to input the title, text, tags, metadata, etc. of articles and videos.

[1820] 2. Send a prediction request:

[1821] The terminal converts the input content data into JSON format and sends it to the server.

[1822] 3. Receiving and displaying prediction results:

[1823] The terminal receives the prediction results and optimization suggestions returned from the server and displays them to the user.

[1824] User operations

[1825] The user accesses the system through a terminal and optimizes the content in the following steps.

[1826] 1. Content Creation:

[1827] Users use their devices to create new article or video content, for example by typing: "How to read the future with AI technology."

[1828] 2. Prediction request:

[1829] The user clicks the button requesting a prediction of the number of views of the content, and transmits data from the terminal to the server.

[1830] 3. Check and correct the predictions:

[1831] Users can check the prediction results and optimization suggestions displayed on their devices. For example, they may see "Expected page views: 1,500 PV" or "Recommended title: 'How to predict the future with AI technology.'" Based on this, users can revise the title and thumbnail and publish the optimized content.

[1832] Example prompt sentence:

[1833] The system collects and preprocesses data in the following way, then analyzes it to predict the number of views of the content. The specific steps are as follows:

[1834] 1. Collect news sites, blogs, and social media posts using web scraping tools and APIs.

[1835] 2. Clean the collected data and convert it into a suitable format for analysis.

[1836] 3. Extract features using natural language processing techniques.

[1837] 4. Train a machine learning model to predict viewership.

[1838] 5. Provide users with real-time prediction results.

[1839] Users can generate optimal titles and descriptions for various content to maximize views. For example, a video title like "How to Read the Future with AI Technology" is expected to attract many viewers according to the predictive model.

[1840] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1841] Step 1: Data collection

[1842] The server collects public data from news sites, blogs, social media posts, and other sources on the internet. To do this, it uses a web scraping tool or API. Specifically, it sends HTTP requests to specified URLs and parses the resulting HTML data to extract usable information. A list of URLs is given as input, and text data is obtained as output.

[1843] Step 2: Data Preprocessing

[1844] The server preprocesses the collected text data by removing HTML tags, normalizing special characters, and filtering irrelevant data. Raw text data is given as input, and clean text data is obtained as output. Regular expressions and text cleaning libraries are used in this process.

[1845] Step 3: Feature extraction

[1846] The server extracts features using the preprocessed text data. To do this, it uses natural language processing (NLP) techniques. Specifically, it uses the Python library Scikit-learn to perform TF-IDF vectorization. Clean text data is given as input, and vectorized feature data is obtained as output.

[1847] Step 4: Train the machine learning model

[1848] The server uses the feature data to train a machine learning model. This uses algorithms such as regression analysis and neural networks. Specifically, it uses the LinearRegression model from Scikit-learn. Vectorized feature data and label data (past view counts) are given as input, and the trained model is obtained as output.

[1849] Step 5: Predicting views

[1850] The server uses the trained machine learning model to predict the number of views for new content data provided by the user. The server receives feature data of the new content as input and obtains the predicted number of views as output.

[1851] Step 6: Optimization suggestions

[1852] The server generates optimal title and thumbnail suggestions based on the prediction results, taking into account past data on similar methods that have generated large numbers of views. The predicted number of views and content feature data are given as input, and the output is an optimized suggestion.

[1853] Step 7: Providing prediction results and optimization suggestions

[1854] The server provides the prediction results and optimization suggestions to the user in real time. This information is displayed to the user through the terminal. The prediction results and optimization suggestions are given as input and displayed to the user as output.

[1855] Step 8: Optimize your content for users

[1856] The user modifies the title and thumbnail of the content based on the provided prediction results and optimization suggestions, which effectively maximizes the number of views of the content. The prediction results and optimization suggestions are given as input, and the optimized content is obtained as output.

[1857] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1858] Overall system overview

[1859] This system collects, preprocesses, and analyzes public data on the Internet, and then predicts the number of views of content based on that data. Furthermore, it incorporates an emotion engine that recognizes user emotions, allowing for predictions that take emotion data into account. This system is primarily comprised of three components: a server, a terminal, and a user.

[1860] Server Processing

[1861] The server has the following functions:

[1862] 1. Data Collection

[1863] The server periodically collects public data from the internet using web scraping tools and APIs.

[1864] The server retrieves text and media data from news sites, blogs, social media, etc. and stores it in storage.

[1865] 2. Data Preprocessing

[1866] The server removes HTML tags from the collected data, normalizes special characters, and converts them into clean text data.

[1867] The server filters out irrelevant data and noise and converts it into a format suitable for analysis.

[1868] 3. Feature Extraction

[1869] The server uses natural language processing (NLP) technology to extract features such as keywords, topics, and sentiment analysis results from the text data.

[1870] The server also performs frame analysis and speech-to-text conversion on video data to extract similar features.

[1871] 4. Analysis by Emotion Engine

[1872] The server uses an emotion engine to recognize emotions from text data about users and content, for example classifying positive, negative, and neutral emotions.

[1873] The server incorporates the emotion data into the pre-processing and feature extraction data.

[1874] 5. Training the predictive model and making predictions

[1875] The server uses the pre-processed and feature-extracted data and the sentiment data to create a training dataset for machine learning.

[1876] The server trains a view prediction model using a regression analysis model or a neural network.

[1877] The server stores the trained model and makes it available for real-time predictions.

[1878] 6. Providing prediction results

[1879] The server generates prediction results and provides them to users in real time, including predicted views, virality predictions, suggestions for optimal presentation and composition, and sentiment-based recommendations.

[1880] Terminal handling

[1881] The terminal provides an interface for users to access the system and input content.

[1882] 1. Content input

[1883] The user enters the title, text, tags, metadata, etc. of the article or video.

[1884] 2. Send a prediction request

[1885] The terminal converts the input content data into JSON format and sends it to the server.

[1886] 3. Check the prediction results

[1887] The terminal receives the prediction results returned from the server and displays them to the user.

[1888] User operations

[1889] The user accesses the system through a terminal and optimizes the content in the following steps.

[1890] 1. Content Creation

[1891] The user uses the device to create new article or video content. For example, the user might enter the following as the title: "Predicting the future with AI" and the text: "It is now possible to predict the future using AI technology. In particular, technology that can predict the number of content views with high accuracy is attracting attention."

[1892] 2. Prediction Request

[1893] The user clicks the button to request a prediction of the number of views of the content, and transmits the data to the server.

[1894] The server performs the analysis and predictions and sends the results back to the device.

[1895] 3. Check and correct the prediction results

[1896] The user checks the prediction results displayed on their device, such as "Expected number of views: 1500 PV," "Recommended title: 'How to read the future with AI technology,'" and "Emotion score: High."

[1897] The user can then modify the title and body of the content based on the prediction results and publish the optimized content.

[1898] Specific examples

[1899] As a concrete example, we show a series of steps in which a user creates a blog post, a server is used to predict the number of views of the post, and optimization is performed taking into account emotional data.

[1900] 1. A user creates a blog post using a device. The title is "How to predict the future with AI technology" and the body is "It is possible to predict the future using AI technology..."

[1901] 2. The user clicks the "View Count Forecast" button, and the device sends the content data to the server.

[1902] 3. The server analyzes the data and extracts emotion data using an emotion engine.

[1903] 4. The server uses the trained prediction model to predict the number of views and generates a prediction result.

[1904] 5. The server sends the prediction results back to the device. For example, "Expected number of views: 1500 PV", "Recommended title: 'How to predict the future with AI'", "Sentiment score: High".

[1905] 6. The user checks the prediction results on their device, changes the title to the recommended title, and publishes the optimized article.

[1906] Through the system of the present invention, users can create effective content based on scientific evidence that takes into account emotional data, thereby supporting the creation of new hit content.

[1907] The processing flow will be explained below.

[1908] Step 1: Data collection

[1909] The server periodically collects public data from the internet using web scraping tools and APIs.

[1910] The server retrieves text and media data from news sites, blogs, social media, etc. and stores it in storage.

[1911] Step 2: Data Preprocessing

[1912] The server removes HTML tags from the collected data, normalizes special characters, and converts them into clean text data.

[1913] The server filters out irrelevant data and noise and converts it into a format suitable for analysis.

[1914] Step 3: Feature extraction

[1915] The server uses natural language processing (NLP) technology to extract features such as keywords, topics, and sentiment analysis results from the text data.

[1916] The server also performs frame analysis and speech-to-text conversion on video data to extract similar features.

[1917] Step 4: Analysis by Emotion Engine

[1918] The server uses an emotion engine to recognize emotions from text data about users and content.

[1919] The server classifies and scores positive, negative, and neutral sentiment within the text data.

[1920] Step 5: Train the predictive model

[1921] The server uses the pre-processed and feature-extracted data and the sentiment data to create a training dataset for machine learning.

[1922] The server trains a view prediction model using a regression analysis model or a neural network.

[1923] The server stores the trained model and makes it available for real-time predictions.

[1924] Step 6: Enter content

[1925] The user uses the device interface to input the title, text, tags, metadata, etc. of the article or video.

[1926] Step 7: Send a prediction request

[1927] When the user clicks the "View Count Prediction" button, the device converts the content data into JSON format and sends it to the server.

[1928] Step 8: Data analysis and prediction

[1929] The server analyzes the received content data and performs pre-processing and feature extraction.

[1930] The server analyzes the emotion of the content using an emotion engine and calculates an emotion score.

[1931] The server uses a trained prediction model to predict the number of views of the content.

[1932] Step 9: Providing prediction results

[1933] The server generates and returns prediction results to the user, including expected views, virality predictions, suggestions for optimal wording and composition, and sentiment scores.

[1934] Step 10: Check and correct the prediction results

[1935] The user checks the prediction results received on their device, which may include, for example, "Expected number of views: 1500 PV," "Recommended title: 'How to read the future with AI technology,'" and "Emotion score: High."

[1936] The user can then modify the title and body of the content based on the prediction results.

[1937] The user publishes the optimized content.

[1938] Example 2

[1939] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1940] Conventional systems that use public data on the Internet to predict the number of views of content are limited to preprocessing and analysis of collected data, and it is difficult to make predictions that take user emotions into account. As a result, prediction accuracy is low and it is difficult to create effective content.

[1941] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for collecting public data on the Internet, means for preprocessing and analyzing the collected data, means for predicting the number of views of content based on the preprocessed and analyzed data, means for extracting and analyzing emotion data using an emotion engine that recognizes the user's emotions, and means for providing the prediction result to the user. This makes it possible to predict the number of views of content with high accuracy, taking the user's emotions into consideration.

[1942] "Public data on the Internet" refers to all information, including text data and media data, that is publicly available on the Internet.

[1943] "Means of collection" refers to the function of regularly obtaining publicly available data on the Internet using web scraping tools and APIs.

[1944] "Preprocessing" refers to the process of removing HTML tags and normalizing special characters from collected data to convert it into clean text data.

[1945] "Means of analysis" refers to the function of analyzing preprocessed data using natural language processing technology, etc., and extracting necessary features and information.

[1946] "Means for predicting the number of content views" refers to a function that uses a machine learning model to predict the number of content views based on pre-processed and analyzed data.

[1947] An "emotion engine" is software or a service for analyzing user emotions, and has the ability to recognize positive, negative, and neutral emotions from text data.

[1948] "Emotion data" is the result data of emotions related to users and content extracted by the emotion engine.

[1949] The "means for providing predicted results" refers to the function of the server to generate predicted results and provide them to the user in real time.

[1950] "Natural language processing technology" is a technology for analyzing text data and extracting features, and includes keyword extraction, topic modeling, sentiment analysis, etc.

[1951] "Content data entered by the user" refers to information such as the title, text, tags, and metadata of articles and videos that the user sends to the system via their terminal.

[1952] "Server" refers to a computer system that centrally handles functions such as data collection, preprocessing, analysis, emotion recognition, and view count prediction.

[1953] "Terminal" refers to the device or application that a User uses to access the System and input Content.

[1954] This system collects, preprocesses, and analyzes public data on the Internet, and predicts the number of views of content based on that data. Furthermore, it combines an emotion engine that recognizes user emotions, allowing for predictions that take emotion data into account.

[1955] Server functions and processing contents

[1956] 1. Data Collection

[1957] The server uses web scraping tools (e.g., BeautifulSoup, Scrapy) and APIs to collect public data from news sites, blogs, social media, etc. For example, article text from news sites and social media posts are stored in a database. The server periodically collects this data and stores it in storage.

[1958] 2. Data Preprocessing

[1959] The server converts the collected data into clean text data by removing HTML tags and normalizing special characters. Specifically, it uses BeautifulSoup to remove unnecessary tags and format the data in a unified format. It also filters out advertisements and irrelevant parts to create a dataset suitable for analysis.

[1960] 3. Feature Extraction

[1961] The server uses natural language processing (NLP) techniques (e.g., spaCy, NLTK) to extract features such as keywords, topics, and sentiment analysis results from text data. For example, it performs topic modeling and sentiment scoring. For video data, it performs frame analysis using OpenCV and speech-to-text conversion using Google Speech-to-Text to extract similar features.

[1962] 4. Analysis by Emotion Engine

[1963] The server analyzes the user's emotions using an emotion engine (e.g., Natural Language Understanding), which classifies positive, negative, and neutral emotions from the text data and integrates the emotion data into a feature dataset.

[1964] 5. Training the predictive model and making predictions

[1965] The server uses the preprocessed and feature-extracted dataset to train a machine learning model (e.g., TensorFlow, PyTorch). The trained predictive model is used to predict the number of views of new content in real time.

[1966] 6. Providing prediction results

[1967] The server uses the trained model to generate predictions and provide them to the user, such as generating JSON data containing predicted views and sentiment-based recommendations, and sending it back to the device.

[1968] User and Device Capabilities

[1969] 1. Content input

[1970] The user uses the device to enter the title, text, tags, and metadata of a new article or video. For example, the title might be "Predicting the future with AI" and the text might be "AI technology makes it possible to predict the future...".

[1971] 2. Send a prediction request

[1972] When the user clicks the "View Count Prediction" button, the device converts the entered content data into JSON format and sends it to the server.

[1973] 3. Check the prediction results

[1974] The device receives the prediction results sent from the server and displays them to the user, such as "Expected number of views: 1500 PV," a recommended title "How to predict the future with AI technology," and an emotion score of "High."

[1975] 4. Modify and optimize your content

[1976] The user checks the prediction results and modifies the content as necessary, for example, changing the title to a recommended title or adjusting the wording of the body text. The optimized content is then published by the user.

[1977] Examples and prompts

[1978] As a concrete example, we show a series of steps in which a user creates a blog post, a server is used to predict the number of views of the post, and optimization is performed taking into account emotional data.

[1979] Title: "How to predict the future with AI technology"

[1980] Body text: "AI technology is making it possible to predict the future. In particular, technology that can predict the number of content views with high accuracy is attracting attention."

[1981] An example prompt might be, "What is the expected number of views for this article?"

[1982] Through this system, users will be able to create effective content based on scientific evidence, helping to create new hit content.

[1983] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1984] Step 1: Data collection

[1985] Input: The server retrieves public data from the internet (e.g., news articles, blog posts, social media posts, etc.) from a specified URL or API.

[1986] What it does: The server collects public data using web scraping tools (e.g., BeautifulSoup, Scrapy) or APIs. For example, it scrapes articles from news sites with BeautifulSoup and retrieves social media posts through APIs.

[1987] Output: The collected raw data is stored in the server's storage. It includes files in various formats, such as text data, image data, and audio data.

[1988] Step 2: Data Preprocessing

[1989] Input: Raw data collected by the server.

[1990] What it does: The server uses BeautifulSoup to strip HTML tags, normalize special characters, and perform noise filtering to remove irrelevant data, such as removing ads and script tags from HTML and resolving encoding issues.

[1991] Output: Clean text data and structured data. For example, ads and irrelevant information are removed and a normalized text file is output.

[1992] Step 3: Feature extraction

[1993] Input: Clean text data.

[1994] Specific operation: The server uses natural language processing (NLP) techniques (e.g., spaCy, NLTK) to extract features such as keywords, topics, and sentiment analysis results. For video data, it performs frame analysis using OpenCV and converts the audio to text using Google Speech-to-Text.

[1995] Output: A dataset of extracted features, including a keyword list and sentiment scores from text data, and important scenes extracted from video frames.

[1996] Step 4: Analysis by Emotion Engine

[1997] Input: Extracted feature dataset.

[1998] Specific operation: The server uses an emotion engine (e.g., Natural Language Understanding) to classify the emotion of each piece of text data and calculate an emotion score, for example, classifying it as "positive," "negative," or "neutral."

[1999] Output: A feature dataset with sentiment data added, e.g., sentiment scores for each text data.

[2000] Step 5: Train the predictive model and make predictions

[2001] Input: Preprocessed and feature extracted dataset.

[2002] What it does: The server uses machine learning models (e.g., TensorFlow, PyTorch) to train regression analysis or neural network models using this data, and then uses the trained models to predict the number of views for new content.

[2003] Output: The trained prediction model and prediction results, such as predicted views and suggested titles for new articles.

[2004] Step 6: Providing prediction results

[2005] Input: JSON data containing prediction results.

[2006] Specific operation: The server generates a prediction result and sends it back to the device. For example, "Expected number of views: 1500 PV" or "Recommended title: 'How to predict the future with AI technology'".

[2007] Output: The prediction result is sent to the device so that the user can check it.

[2008] Step 7: User input and prediction request

[2009] Input: User-generated content data (article title, body, tags, metadata, etc.).

[2010] Specific operation: The user inputs content through the terminal and clicks the "View Count Prediction" button. The terminal converts the content data into JSON format and sends it to the server.

[2011] Output: The content data sent to the server.

[2012] Step 8: Check the prediction results and modify the content

[2013] Input: Prediction results sent from the server.

[2014] Specific operation: The user checks the prediction results on the device and modifies the content as necessary. For example, the title can be changed to "How to read the future with AI technology" based on the prediction results.

[2015] Output: The optimized content is published.

[2016] As described above, data input, data processing, data calculation, and output are performed in each processing step, enabling users to create effective content by utilizing highly accurate prediction results and emotion data.

[2017] (Application example 2)

[2018] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[2019] Conventional content view count prediction systems simply predicted the number of views without considering user emotions, making it difficult to optimize content more effectively. In particular, in the advertising field, where emotion-based optimization is required, there was a lack of functionality to suggest optimal titles and content.

[2020] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[2021] In this invention, the server includes means for collecting public data on the Internet, means for preprocessing and analyzing the collected data, means for predicting the number of views of content based on the preprocessed and analyzed data, means for recognizing user emotions using a sentiment analysis engine and making predictions taking the emotional data into consideration, means for providing the prediction results and optimization suggestions to the user, and means for generating optimal titles and content revision suggestions for advertising content, thereby enabling the prediction and optimization of the number of views of advertising content that reflects user emotions.

[2022] "Public data on the internet" refers to information that is freely accessible on websites, blogs, social media platforms, etc.

[2023] "Means of collection" refers to the process of using web scraping tools or APIs to obtain public data on a server.

[2024] "Preprocessing" refers to the process of removing unnecessary HTML tags from collected data, normalizing special characters, and converting the data into a format suitable for analysis.

[2025] "Means of analysis" refers to the process of analyzing data using natural language processing techniques and extracting features such as keywords, topics, and sentiment.

[2026] "Method for predicting content views" refers to the process of using a machine learning model to predict content views based on pre-processed and analyzed data.

[2027] "Sentiment analysis engine" refers to an algorithm that recognizes positive, negative, and neutral emotions from text data entered by the user.

[2028] "Means for making predictions taking into account emotional data" refers to a process of incorporating emotional data into a view count prediction model to make view count predictions that take into account the influence of emotions.

[2029] "Means for providing prediction results and optimization suggestions to users" refers to a process for providing users with prediction results as well as suggestions for more effective content expression and configuration.

[2030] "Means for generating optimal titles and suggested content modifications for advertising content" refers to the process of generating effective titles and suggested text modifications for advertising content based on prediction results and emotional data.

[2031] The present invention relates to a system that collects, preprocesses, and analyzes public data on the Internet and predicts the number of views of content based on that data. This system, combined with a sentiment analysis engine, performs advanced predictions that take emotional data into account and provides optimal titles and content revisions, particularly in the advertising field. A specific embodiment of this system is described below.

[2032] Server Processing

[2033] The server has the following main functions:

[2034] 1. Data Collection

[2035] The server periodically collects public data from the internet using web scraping tools and APIs.

[2036] For example, text and media data is acquired from news sites, blogs, and social media and stored in a database.

[2037] 2. Data Preprocessing

[2038] The server removes HTML tags from the collected data, normalizes special characters, and converts them into clean text data.

[2039] It also filters out irrelevant data and noise and converts it into a format suitable for analysis.

[2040] 3. Feature Extraction

[2041] The server uses natural language processing technology (e.g., spaCy or NLTK) to extract features such as keywords, topics, and sentiment analysis results from the text data.

[2042] In the case of video data, frame analysis and speech-to-text conversion are performed to extract similar features.

[2043] 4. Analysis using a sentiment analysis engine

[2044] The server uses a sentiment analysis engine (such as VADER) to classify the user's sentiment from the text data.

[2045] The classified sentiment data (positive, negative, neutral) is incorporated into the preprocessing and feature extraction data.

[2046] 5. Training the predictive model and making predictions

[2047] The server uses the preprocessed and feature-extracted data and sentiment data to train a machine learning model (e.g., Scikit-learn or TensorFlow).

[2048] The trained model is used to predict the number of views of the content.

[2049] 6. Providing prediction results and optimization suggestions

[2050] The server generates and provides the user with optimal representation and composition suggestions along with the prediction results.

[2051] Specifically, it generates optimal titles and suggested content revisions for advertising content, reflecting user emotional data.

[2052] Terminal handling

[2053] The terminal provides an interface for users to access the system and input content.

[2054] 1. Content input

[2055] Users enter the title, text, tags, metadata, etc. of the article or video.

[2056] 2. Send a prediction request

[2057] The terminal converts the input content data into JSON format and sends it to the server.

[2058] 3. Check the prediction results and optimization suggestions

[2059] The terminal receives the prediction results and optimization suggestions returned from the server and displays them to the user.

[2060] User operations

[2061] A user accesses the system through a terminal and optimizes advertising content by following the steps below.

[2062] 1. Content Creation

[2063] The user uses the device to create new advertising content, for example, by typing "We've launched a new smartwatch. Learn more!"

[2064] 2. Prediction Request

[2065] The user clicks the button to request a prediction of the number of views of the content, and transmits the data to the server.

[2066] 3. Check and modify the prediction results and optimization suggestions

[2067] Users can view the prediction results and optimization suggestions displayed on their device, such as "Expected page views: 3,000 PV," "Recommended title: 'Limited sale! Get it now,'" and "Sentiment score: High."

[2068] Users optimize and publish content based on prediction results and suggestions.

[2069] Specific examples

[2070] When a user submits an ad to the system saying, "We've released a new smartwatch. Click here for details!", the system performs sentiment analysis and predicts the number of views. The predicted results are displayed as "Sentiment score: Medium," "Expected number of views: 2000 PV," and "Recommended title: 'Limited time only! Check out the details of the limited edition smartwatch.'" The user can then change the displayed title to optimize the ad.

[2071] Prompt Sentence Examples

[2072] Prompt statement:

[2073] The ad data entered by the user is listed below. Please perform sentiment analysis and respond with a view prediction and optimal ad modification suggestions.

[2074] Ad title: New smartwatch released

[2075] Ad text: We've launched a new smartwatch. Learn more here!

[2076] Tags: smartwatch sale

[2077] Expected output:

[2078] Estimated number of views: XXXX PV

[2079] Recommended title: 'Now only! Check out the details of the limited edition smartwatch'

[2080] Sentiment score: High / Medium / Low

[2081] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[2082] Step 1:

[2083] Data collection (server)

[2084] Input: Public data from the internet (e.g., news articles, blog posts, social media posts).

[2085] What it does: The server periodically collects this public data using web scraping tools and APIs and stores it in a database.

[2086] Output: The raw data collected.

[2087] Step 2:

[2088] Data preprocessing (server)

[2089] Input: The raw data collected.

[2090] What it does: The server strips HTML tags, normalizes special characters, filters irrelevant data and noise, and generates clean text data.

[2091] Output: Preprocessed and clean text data.

[2092] Step 3:

[2093] Feature extraction (server)

[2094] Input: Preprocessed clean text data.

[2095] Specific operation: The server uses natural language processing techniques (e.g., spaCy or NLTK) to extract features such as keywords, topics, and sentiment analysis results from text data. For video data, it also performs frame analysis and speech-to-text conversion.

[2096] Output: Feature-extracted data (keywords, topics, sentiment scores, etc.).

[2097] Step 4:

[2098] Sentiment analysis (server)

[2099] Input: Feature extracted data.

[2100] What happens: The server uses a sentiment analysis engine (e.g., VADER) to classify specific sentiments (positive, negative, neutral) and calculate a sentiment score.

[2101] Output: Data with sentiment scores.

[2102] Step 5:

[2103] Training and predicting the predictive model (server)

[2104] Input: Preprocessed and sentiment-scored data.

[2105] What it does: The server uses a machine learning model (e.g., Scikit-learn or TensorFlow) to train a model to predict the number of content views and makes the prediction.

[2106] Output: View count prediction results.

[2107] Step 6:

[2108] Generation of prediction results and optimization proposals (server)

[2109] Input: View count prediction results and sentiment score.

[2110] Specific operation: Based on the prediction results, the server generates optimal ad titles and content revision suggestions and provides them to the user.

[2111] Output: Forecast results with optimization suggestions.

[2112] Step 7:

[2113] Content input (terminal)

[2114] Input: Ad content (e.g. title, body, tags).

[2115] Specific actions: The user uses the device to enter data into the content input interface and clicks the submit button.

[2116] Output: The input content data is sent to the server.

[2117] Step 8:

[2118] Sending a prediction request (device)

[2119] Input: The content data entered.

[2120] Specific operation: The terminal converts the content data into JSON format and sends it to the server.

[2121] Output: The content data sent to the server.

[2122] Step 9:

[2123] Check prediction results and optimization suggestions (device)

[2124] Input: Prediction results and optimization suggestions returned by the server.

[2125] Specific operation: The device receives the prediction results and optimization suggestions returned from the server and displays them to the user.

[2126] Output: Prediction results and optimization suggestions displayed to the user.

[2127] Step 10:

[2128] Content optimization and publishing (user)

[2129] Input: Prediction results and optimization suggestions.

[2130] What it does: The user modifies the content based on the displayed predictions and suggestions, then clicks the publish button to publish the optimized content.

[2131] Output: The published optimized content.

[2132] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[2133] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[2134] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[2135] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[2136] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[2137] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[2138] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[2139] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[2140] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[2141] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[2142] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[2143] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[2144] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[2145] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[2146] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[2147] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[2148] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[2149] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[2150] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[2151] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[2152] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[2153] The following is further disclosed regarding the above embodiment.

[2154] (Claim 1)

[2155] A means for collecting public data on the Internet;

[2156] a means for preprocessing and analyzing the collected data;

[2157] A means for predicting the number of views of content based on the preprocessed and analyzed data;

[2158] a means for providing the prediction result to a user;

[2159] A system including:

[2160] (Claim 2)

[2161] 10. The system of claim 1, further comprising means for extracting content features using natural language processing techniques.

[2162] (Claim 3)

[2163] means for transmitting content data input by a user to a server;

[2164] means for returning and displaying the prediction results from the server to the user;

[2165] 10. The system of claim 1, further comprising:

[2166] "Example 1"

[2167] (Claim 1)

[2168] A means for collecting public data on the Internet;

[2169] a means for pre-processing the collected data, including removing HTML tags, normalizing special characters, and filtering irrelevant data;

[2170] A means of tokenizing the preprocessed data and converting it into a format suitable for analysis; and

[2171] A means of extracting features such as keywords, topics, and sentiment analysis results from text data using natural language processing technology;

[2172] A means of training a machine learning model based on the analyzed data to predict the number of views of the content; and

[2173] a means for providing the prediction results to a user in real time;

[2174] A system including:

[2175] (Claim 2)

[2176] Users enter content such as article or video titles, text, tags, and metadata,

[2177] A means for converting input content data into JSON format and transmitting the converted data to a server;

[2178] means for returning and displaying the prediction results from the server to the user;

[2179] 10. The system of claim 1, further comprising:

[2180] (Claim 3)

[2181] We perform frame analysis and audio-to-text conversion on video data,

[2182] further comprising means for performing specific analysis and feature extraction;

[2183] 10. The system of claim 1.

[2184] "Application Example 1"

[2185] New Claims

[2186] (Claim 1)

[2187] A means for collecting public data on the Internet;

[2188] a means for preprocessing and analyzing the collected data;

[2189] A means for predicting the number of views of content based on the preprocessed and analyzed data;

[2190] A way to optimize content titles and thumbnails based on predicted views;

[2191] means for providing prediction results and optimization suggestions to a user;

[2192] A system including:

[2193] (Claim 2)

[2194] 10. The system of claim 1, further comprising means for extracting content features using natural language processing techniques.

[2195] (Claim 3)

[2196] means for transmitting content data input by a user to a server;

[2197] means for returning and displaying the prediction results and optimization suggestions from the server to the user;

[2198] 10. The system of claim 1, further comprising:

[2199] "Example 2: Combining Emotion Engines"

[2200] (Claim 1)

[2201] A means for collecting public data on the Internet;

[2202] a means for preprocessing and analyzing the collected data;

[2203] A means for predicting the number of views of content based on the preprocessed and analyzed data;

[2204] means for extracting and analyzing emotion data using an emotion engine that recognizes the user's emotions;

[2205] a means for providing the prediction result to a user;

[2206] A system including:

[2207] (Claim 2)

[2208] 10. The system of claim 1, further comprising means for extracting content features using natural language processing techniques.

[2209] (Claim 3)

[2210] means for transmitting content data input by a user to a server;

[2211] means for returning and displaying the prediction results from the server to the user;

[2212] 10. The system of claim 1, further comprising:

[2213] "Application example 2 when combining emotion engines"

[2214] (Claim 1)

[2215] A means for collecting public data on the Internet;

[2216] a means for preprocessing and analyzing the collected data;

[2217] A means for predicting the number of views of content based on the preprocessed and analyzed data;

[2218] means for recognizing and predicting user emotions using a sentiment analysis engine and taking into account the sentiment data;

[2219] a means for providing prediction results and optimization suggestions to a user;

[2220] A means for generating optimal titles and content revisions for advertising content;

[2221] A system including:

[2222] (Claim 2)

[2223] 10. The system of claim 1, further comprising means for extracting content features using natural language processing techniques.

[2224] (Claim 3)

[2225] means for transmitting advertising content data input by a user to a server;

[2226] a means for returning and displaying the prediction results and optimization suggestions from the server to the user;

[2227] 10. The system of claim 1, further comprising: [Explanation of symbols]

[2228] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. A means for collecting public data on the Internet; a means for preprocessing and analyzing the collected data; A means for predicting the number of views of content based on the preprocessed and analyzed data; a means for providing the prediction result to a user; A system including:

2. The system of claim 1 , further comprising means for extracting content features using natural language processing techniques.

3. means for transmitting content data input by a user to a server; means for returning and displaying the prediction results from the server to the user; The system of claim 1 further comprising:

Images