System

Abstract

A system is provided.SOLUTION: A system, comprising: means for collecting historical event data; means for training a model based on the collected historical event data; means for collecting current trend data; means for predicting future popularity from the current trend data using the trained model; and means for recommending an advertisement destination based on the predicted popularity.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] In the modern advertising industry, certain events or content can rapidly attract attention. However, if the timing for placing an advertisement is missed, a business opportunity can be missed. To solve this problem, a system is needed that utilizes past event data and current trend data to predict future popularity and recommend advertising at the appropriate time. The system of the present invention aims to optimize advertising operations by effectively performing such predictions and recommendations. [Means for solving the problem]

[0005] The present invention solves the above problems by providing a system including the following means.

[0006] a means for collecting historical event data;

[0007] A means of training a model based on collected historical event data; and

[0008] a means of collecting current trend data;

[0009] A means for predicting future popularity from current trend data using the trained model;

[0010] A means for recommending advertising destinations based on the predicted popularity;

[0011] It is a system including:

[0012] The system also includes a means for preprocessing collected past event data and extracting features and target variables, and a means for training a model using the preprocessed data. The system also includes a means for preprocessing collected current trend data and extracting features, and a means for predicting future popularity based on the preprocessed current trend data, thereby improving prediction accuracy. The system is thus capable of effectively recommending advertising destinations to users.

[0013] "Past event data" refers to past information about a specific event or content, including data about the popularity and attention of the event, such as viewership ratings, number of participants, number of followers on social media, and number of related news articles.

[0014] "Training a model" refers to the process of using a machine learning algorithm to learn a predictive model based on collected data, and optimizing the model parameters using the data features and target variables.

[0015] "Current trend data" refers to data that indicates how much attention a particular event or piece of content is currently receiving, and includes data such as the number of social media posts, news articles, and search engine queries.

[0016] "Predicting future popularity" refers to the process of predicting how much attention or popularity an event or content will attract in the future based on current trend data, and calculating a popularity score using a predictive model.

[0017] "Recommending advertising destinations" refers to the process of proposing the optimal timing and location for advertising based on predicted future popularity data, making recommendations to maximize the effectiveness of advertising.

[0018] "Preprocessing" refers to the process of converting data into a format suitable for learning and prediction, and includes data cleansing, feature extraction, and target variable selection.

[0019] "Features" refer to the attributes or variables of the data used to train the model and are used as input values ​​for the predictive model.

[0020] A "target variable" refers to the desired value that a model predicts and is used as the output value of a predictive model.

[0021] "SNS" is an abbreviation for social networking service, and refers to a platform where users can communicate and share information online. [Brief explanation of the drawings]

[0022] [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

[0023] 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.

[0024] First, the terms used in the following description will be explained.

[0025] 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).

[0026] 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.

[0027] 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.

[0028] 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.

[0029] 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."

[0030] [First embodiment]

[0031] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0032] 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.

[0033] 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).

[0034] 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.

[0035] 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.

[0036] 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.

[0037] 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.

[0038] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0039] 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.

[0040] 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.

[0041] 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.

[0042] 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."

[0043] The present invention relates to a system for recommending advertisement placements, and includes processes for collecting past event data, training a model, collecting current trend data, predicting future popularity based on that data, and recommending advertisement placements. The program processing of this system is explained below in natural language.

[0044] Program Overview

[0045] The server retrieves the current trend

[0046] The server sends requests to social media and news APIs to retrieve current topics. This data, which contains information about specific topics or events, is returned in JSON format. The server analyzes this trend data to extract key keywords and temporal features.

[0047] The server collects and analyzes past event data

[0048] The server uses keywords extracted from current trends to collect past event data from a database or another API. This past data includes information about the popularity and attention of the event, such as viewer ratings, number of participants, number of social media posts, and number of related news articles. The server preprocesses this data and extracts features and target variables.

[0049] The server trains the model

[0050] The server uses the preprocessed historical event data to train a machine learning algorithm (e.g., a linear regression model). During this training process, features are input, popularity scores are output, and the model learns the relationship between them.

[0051] The server predicts future popularity

[0052] The server uses the trained model to predict future popularity from current trend data, allowing it to predict which events and content will attract attention in the future. The server adds the prediction results to the trend data and assigns a predicted popularity score to each trend.

[0053] The server recommends advertising

[0054] The server selects trends suitable for advertising based on the predicted popularity, prioritizes trends with high predicted popularity, and recommends the top trends that are expected to be particularly effective as advertising targets.

[0055] Specific examples

[0056] For example, if there is a sudden surge in excitement for the 2023 Basketball World Cup, the system would work as follows:

[0057] 1. The server collects current topics about the Basketball World Cup from social media and news APIs. For example, it detects that the hashtag "BasketballWorldCup2023" is frequently used.

[0058] 2. The server collects past sporting event data (e.g., the last Basketball World Cup, the Olympics) and preprocesses their viewership and attendance figures.

[0059] 3. The server uses this historical data to train a machine learning model (e.g., a linear regression model).

[0060] 4. The server uses the trained model to predict future popularity based on current trend data. The prediction reveals that the Basketball World Cup 2023 will peak next weekend.

[0061] 5. Based on the prediction results, the server recommends "Basketball World Cup 2023" as an advertising destination. It assigns a score indicating the applicability and priority of the advertising destination and recommends the optimal timing for advertising.

[0062] 6. After the ads are actually placed, the server collects the results and feeds them back into the model to improve prediction accuracy.

[0063] In this way, the system of the present invention can respond to trending content in real time and effectively recommend advertising, allowing advertisers to maximize their business opportunities without missing out.

[0064] The processing flow will be explained below.

[0065] Program processing steps

[0066] Step 1:

[0067] The server sends a request to a social networking or news API, such as the Twitter API or Google News API, to get the most popular topics at the moment, which allows you to get current trending data.

[0068] Step 2:

[0069] The server receives the API response and retrieves the trending data in JSON format, including specific keywords, topics, hashtags, etc.

[0070] Step 3:

[0071] The server analyzes the trend data and extracts keywords and temporal features (e.g., event dates and times, related hashtags) that are likely to be relevant to advertising. This data is used for subsequent model training.

[0072] Step 4:

[0073] The server then collects past event data from a database or other APIs based on the extracted keywords. For example, trending keywords related to the Basketball World Cup are used to collect data on similar past sporting events.

[0074] Step 5:

[0075] The server preprocesses past event data collected by the server. Specifically, it extracts features and target variables from data such as viewership, number of participants, and number of posts on social media. This creates a dataset in a format suitable for learning.

[0076] Step 6:

[0077] The server uses the preprocessed data to train a machine learning algorithm (e.g., a linear regression model). Through this training process, the model learns the relationship between features and popularity scores.

[0078] Step 7:

[0079] The server uses the trained model to predict future popularity based on current trend data. For example, it uses the current trend data of the Basketball World Cup as input and predicts future viewership and the likelihood of it becoming a hot topic on social media.

[0080] Step 8:

[0081] The server adds the predictions to the trend data and gives each trend a predicted popularity score, which reveals which trends are most suitable for advertising.

[0082] Step 9:

[0083] The server selects trends suitable for advertising based on predicted popularity, prioritizes trends with high predicted popularity, and recommends the top trends that are expected to be particularly effective as advertising targets.

[0084] Step 10:

[0085] The server notifies the user of the ad placement destinations. Specifically, it presents the user with a list of recommended ad placement destinations, along with details of the timing and location.

[0086] Step 11:

[0087] After the ad is placed, the server collects actual popularity data, such as viewer ratings and social media reactions after the ad is placed.

[0088] Step 12:

[0089] The server analyzes the difference between the actual popularity data and the predictions to identify areas for improvement in the model, providing feedback to improve the accuracy of the prediction model.

[0090] Step 13:

[0091] The server retrains the model using actual popularity data to improve prediction accuracy, enabling more accurate predictions for future ad placements.

[0092] Through the steps described above, the system of the present invention can respond to rapidly trending content in real time and effectively recommend advertising placements.

[0093] Example 1

[0094] 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."

[0095] Conventional advertising systems do not fully utilize past data and lack real-time trend analysis. As a result, they are unable to optimally select advertising destinations, resulting in reduced advertising effectiveness. Furthermore, they lack a feedback function to maintain and improve the accuracy of their predictive models, making it difficult to maximize advertising effectiveness over the long term.

[0096] 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.

[0097] In this invention, the server includes means for collecting past data, means for training a model based on the collected past data, means for collecting current trend data, means for predicting future popularity from the current trend data using the trained model, means for recommending advertising destinations based on the predicted popularity, means for acquiring and analyzing trend data from an API, means for preprocessing past data and extracting features and target variables, means for assigning popularity scores to each trend, and means for collecting actual results of advertising as feedback and improving the model. This enables advertisers to respond to rapidly trending content in real time and place effective advertisements.

[0098] "Past data" refers to information about past events, including viewership ratings, number of participants, number of posts on social media, and number of related news articles.

[0099] "Training a model" means using collected data to build a computational model using machine learning algorithms to learn patterns from the data and make predictions.

[0100] "Current trend data" is real-time information related to specific topics or events obtained from social media and news sites.

[0101] "Predicting future popularity" means using a trained model to predict the popularity of future events or topics based on current trend data.

[0102] "Recommending advertising destinations" means suggesting which events or topics you should advertise on based on predicted future popularity information.

[0103] "Obtaining and analyzing trend data from an API" means using an application programming interface (API) to obtain data from social media and news sites, and analyzing that data to extract useful information.

[0104] A "feature" is a specific attribute or characteristic of data that is used as input for a machine learning model.

[0105] A "target variable" is a value or category that is predicted as the output of a machine learning model.

[0106] A "Popularity Score" is a numerical representation of how much attention a particular event or topic will receive in the future.

[0107] "Feedback" refers to data collected from the results of actual advertising placements (e.g., click-through rate, conversion rate) and used to improve the performance of the model.

[0108] The present invention relates to a system for recommending advertisements, which trains a machine learning model based on past data and current trend data, predicts future popularity, and recommends advertisement placements. Specific embodiments for implementing this system are described below.

[0109] System configuration and technologies used

[0110] The server sends requests to the APIs of social media platforms and news services (e.g., Twitter API, Google News API) to obtain current trending data, which is returned in JSON format and parsed by the server to extract information about specific topics or events.

[0111] The server uses the extracted keywords to collect past data from a database or another API (e.g., sports database API, Google Trends API). The collected data includes viewership, number of participants, number of social media posts, number of related news articles, etc. The server preprocesses this data and extracts features and target variables. Specifically, it imputes missing values ​​and normalizes the data.

[0112] The server uses the preprocessed data to train a model using a machine learning algorithm (e.g., linear regression model or random forest). This training process uses the Scikit-learn library, with features as input and popularity scores as output. The dataset is split into training data and test data, and the performance of the model is evaluated.

[0113] The server uses the trained model to predict future popularity based on current trend data. Specifically, it inputs data from the "Basketball World Cup 2023" obtained from the Twitter API into the model and calculates the predicted level of attention for the following weekend. The output is a popularity score, which is used to evaluate future trends.

[0114] The server recommends optimal advertising destinations based on the predicted popularity score. For example, if interest in the "Basketball World Cup 2023" is predicted to be at its highest next weekend, the server will recommend this trend as an advertising destination. It also provides specific advertising strategies and timing to advertisers.

[0115] The server collects the results data (e.g., click-through rate, conversion rate) after the ads are actually placed and uses them as feedback to improve the accuracy of the model. The results data are used to retrain the model, maximizing the effectiveness of advertising in the long term.

[0116] Examples of specific examples and prompts

[0117] For example, if the 2023 Basketball World Cup is trending rapidly on social media, the system would work like this:

[0118] 1. The server collects current trend data related to "BasketballWorldCup2023" from the Twitter API and Google News API, and checks the frequency of the extracted keyword "BasketballWorldCup2023."

[0119] 2. The server collects viewership data and participant numbers from the previous Basketball World Cup from the sports database API, completes the missing data, and extracts it as features.

[0120] 3. The server uses Scikit-learn to train a linear regression model based on past data and build a model to predict the popularity of events.

[0121] 4. The server uses the trained model to predict future interest in the Basketball World Cup 2023 and estimates the peak interest for the following weekend.

[0122] 5. Based on the prediction results, the server recommends "Basketball World Cup 2023" as an advertising destination and proposes specific advertising timing and strategies to the advertiser.

[0123] 6. The server collects post-advertising result data (click-through rate and conversion rate) and retrains the predictive model to improve accuracy.

[0124] Example prompt sentence:

[0125] "Collect current social media trends related to the 2023 Basketball World Cup, use past sporting event data to predict future popularity, and recommend optimal advertising destinations. Specifically, retrieve trend data from the Twitter API, train a linear regression model using Scikit-learn, and propose an advertising strategy based on the prediction results."

[0126] In this way, the system of the present invention provides advertisers with a means to effectively place advertisements in response to rapidly trending content in real time.

[0127] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0128] Step 1:

[0129] The server collects current trend data. It sends requests to APIs of social media platforms and news services (e.g., Twitter API, Google News API) to retrieve data about current topics. The input is the API request, and the output is trend data in JSON format. The server parses this data to extract key information about specific topics or events, such as the frequency of occurrence of the hashtag "BasketballWorldCup2023."

[0130] Step 2:

[0131] The server collects past data. Using the keywords extracted in step 1, the server retrieves past related event data from a database or another API (e.g., sports database API, Google Trends API). This includes viewership, number of participants, number of social media posts, number of related news articles, etc. The input is keywords related to past events, and the output is a dataset of related past events. The server parses this dataset to extract the necessary information.

[0132] Step 3:

[0133] The server preprocesses the data. The server normalizes the past data collected in step 2, fills in missing values, and formats it as features and target variables. For example, it formats viewer ratings and number of posts as numerical data. The input is raw data of past events, and the output is a preprocessed dataset.

[0134] Step 4:

[0135] The server trains the model. The server uses the preprocessed data to train the model using a machine learning algorithm (e.g., linear regression model or random forest). For example, it uses the Scikit-learn library. The input is the preprocessed dataset, and the output is the trained predictive model. The model is configured to input features and output a popularity score.

[0136] Step 5:

[0137] The server predicts future popularity. The server uses the trained model to predict future popularity from the current trend data obtained in step 1. The input is the current trend data and the trained model, and the output is the future popularity score. For example, the server predicts that the hashtag "BasketballWorldCup2023" will peak next weekend.

[0138] Step 6:

[0139] The server recommends advertising destinations. The server selects the optimal advertising destination based on the popularity score predicted in step 5 and recommends advertising to the advertiser. The input is the future popularity score, and the output is the recommended advertising destination. Specifically, the server recommends "Basketball World Cup 2023" as an advertising destination and suggests advertising strategies and timing.

[0140] Step 7:

[0141] The server collects feedback and improves the model. The server collects results after the ads are actually placed (e.g., click-through rate, conversion rate), and uses that data to retrain the model and improve its accuracy. The input is the results data after the ads are placed, and the output is an improved model. This makes it possible to maximize advertising effectiveness in the long term.

[0142] (Application example 1)

[0143] 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."

[0144] Conventional advertising systems statically select targets based on past data and are unable to respond to real-time trend changes, making it difficult to optimize advertising effectiveness. Furthermore, there is no feedback based on post-advertising results to improve model accuracy, meaning long-term improvements in advertising effectiveness cannot be expected. Furthermore, there is a lack of a way for users to easily evaluate advertising effectiveness and reflect that in their next ad placement.

[0145] 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.

[0146] In this invention, the server includes means for collecting past event data, means for training a model based on the collected past event data, means for collecting current trend data, means for predicting future popularity from the current trend data using the trained model, means for recommending advertising destinations based on the predicted popularity, means for users to provide feedback on the results of their advertising placements, and means for improving the accuracy of the model using the feedback results. This enables rapid response to trend fluctuations in real time and optimization of advertising effectiveness. Furthermore, improving the accuracy of the model based on user feedback can also be expected to improve long-term advertising effectiveness. Furthermore, users can easily evaluate advertising effectiveness and reflect it in their next advertising placements.

[0147] "Past event data" refers to information related to events that have taken place in the past, including viewership ratings, number of participants, number of posts on social media, number of related news articles, and the like.

[0148] "Training a model" is the process of using machine learning algorithms to learn relationships between collected data.

[0149] "Current trending data" is data on information and topics that are currently trending, collected from social media, news, and other sources.

[0150] "Predicting future popularity" is the process of using a trained model to infer what will be popular and trending in the future based on current trend data.

[0151] "Recommending advertising locations" means suggesting the most effective locations and timing for advertising based on predicted popularity.

[0152] "Providing feedback" means that the user reports to the server the results and effects obtained after placing an advertisement.

[0153] "Improving the accuracy of the model" means adjusting the parameters and algorithms of the predictive model based on the feedback results to increase the accuracy of the next prediction.

[0154] This invention relates to a system for recommending advertising, which consists of the following steps: First, a server collects past event data and trains a machine learning model based on this data. Next, the server collects current trend data and uses the trained model to predict future popularity. Based on the prediction results, the system recommends optimal advertising destinations to users.

[0155] Program processing

[0156] Data Collection Module

[0157] The server collects current trend data via social media and news APIs. Specifically, it uses Python to send requests to the Twitter API and Google News API, and retrieves JSON-formatted data including trending keywords and the number of posts.

[0158] Hardware used: Server, smartphone

[0159] Software used: Python, Twitter API, Google News API

[0160] Historical Data Analysis Module

[0161] The server collects past event data from a database or another API, and performs preprocessing to extract features and target variables. In this process, metrics such as viewership, number of participants, and number of posts on social media are used, and the data is organized using Python's Scikit-learn.

[0162] Hardware used: Server

[0163] Software used: Python, Scikit-learn

[0164] Model Training Module

[0165] The server uses the preprocessed data to train a machine learning algorithm (e.g., a linear regression model), which learns relationships in the data and generates a model to predict future popularity.

[0166] Hardware used: Server

[0167] Software used: Python, Scikit-learn

[0168] Popularity Prediction Module

[0169] The server uses the trained model to predict future popularity based on current trend data, a process also performed using Python.

[0170] Hardware used: Server

[0171] Software used: Python, Scikit-learn

[0172] Recommendation Engine

[0173] The server recommends the most effective advertising destinations based on predicted popularity, allowing users to know effective advertising destinations in real time.

[0174] Hardware used: Server

[0175] Software used: Python, Flask

[0176] Feedback Module

[0177] Users provide feedback on the results of their ad placements, and the data is sent to the server, where it is used to retrain the model, improving its prediction accuracy.

[0178] Hardware used: Smartphone

[0179] Software used: React Native, Firebase

[0180] Specific examples

[0181] As a specific application example, we will explain the process in which a server collects the latest trend data and predicts future popularity. For example, the server inputs the following prompt sentence into the generative AI model:

[0182] Gather the latest trends and predict future popularity. Use past event data to identify hot topics for the upcoming weekend and suggest them for advertising.

[0183] Through this process, the server can quickly respond to trend fluctuations in real time and optimize advertising effectiveness. It also has a feedback function to improve the accuracy of the model, making it a system that can achieve long-term improvements in advertising effectiveness.

[0184] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0185] Step 1:

[0186] The server collects past event data. Specifically, the server retrieves data related to past events, such as viewer ratings, number of participants, and number of posts on social media, from a database or API. After collection, this data is saved in JSON format and used for subsequent processing.

[0187] Input: Historical event data from a database or API

[0188] Output: Historical event data stored in JSON format

[0189] Step 2:

[0190] The server preprocesses the collected past event data and extracts features and target variables. Specifically, it performs missing value imputation, noise removal, and scaling of the data, and then extracts features (e.g., viewership rate, number of participants, number of social media posts) and target variables. Python's Scikit-learn is used.

[0191] Input: Historical event data stored in JSON format

[0192] Output: Preprocessed features and target variable

[0193] Step 3:

[0194] The server trains the model using the preprocessed data. Specifically, it uses a machine learning algorithm such as a linear regression model, inputs features, and outputs target variables. Python's Scikit-learn is used.

[0195] Input: Preprocessed features and target variables

[0196] Output: A trained machine learning model

[0197] Step 4:

[0198] The server collects current trend data from social media and news APIs. Specifically, it sends an API request and receives data in JSON format containing the latest topics and keywords. For example, it uses the Twitter API or Google News API.

[0199] Input: Trend data from Twitter API and Google News API

[0200] Output: Current trend data saved in JSON format

[0201] Step 5:

[0202] The server preprocesses the collected current trend data and extracts features. Specifically, it performs noise removal and normalization, and extracts features (e.g., topics, keywords, and number of posts). It uses Python's Scikit-learn.

[0203] Input: Current trend data stored in JSON format

[0204] Output: Preprocessed features

[0205] Step 6:

[0206] The server uses the trained model to predict future popularity based on current trend data. Specifically, the preprocessed features are input into the model to predict future popularity scores.

[0207] Input: Preprocessed features, trained model

[0208] Output: Predicted future popularity scores

[0209] Step 7:

[0210] The server recommends advertising destinations based on the predicted popularity. Specifically, it prioritizes trends with high popularity scores and suggests them as advertising destinations. Users receive this information through a smartphone application.

[0211] Input: Predicted future popularity scores

[0212] Output: Recommended ad placements

[0213] Step 8:

[0214] Users provide feedback on the results of their ads after placing them. Specifically, they use their smartphones to input advertising effectiveness measurement data (click rates, conversion rates, etc.) and send it to the server.

[0215] Input: Post-advertising effectiveness measurement data

[0216] Output: Feedback data sent to the server

[0217] Step 9:

[0218] The server uses the feedback data to improve the accuracy of the model. Specifically, it uses the feedback data as retraining data to adjust the model parameters and algorithms, thereby improving the accuracy of the next prediction.

[0219] Input: Feedback data

[0220] Output: A trained model with improved accuracy

[0221] 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.

[0222] The present invention combines an advertising recommendation system with an emotion engine that recognizes user emotions, and in addition to the process of collecting past event data, training a model based on this, collecting current trend data to predict future popularity, and recommending advertising destinations, the system utilizes user emotion data to achieve more accurate advertising placement. The program processing of this system is explained below in natural language.

[0223] Program Overview

[0224] The server retrieves the current trend

[0225] The server sends requests to social media and news APIs to retrieve current topics. This data, which contains information about specific topics or events, is returned in JSON format. The server analyzes this trend data to extract key keywords and temporal features.

[0226] The server collects and analyzes past event data

[0227] The server uses keywords extracted from current trends to collect past event data from a database or another API. This past data includes information about the popularity and attention of the event, such as viewer ratings, number of participants, number of social media posts, and number of related news articles. The server preprocesses this data and extracts features and target variables.

[0228] The server trains the model

[0229] The server uses the preprocessed historical event data to train a machine learning algorithm (e.g., a linear regression model). During this training process, features are input, popularity scores are output, and the model learns the relationship between them.

[0230] The server predicts future popularity

[0231] The server uses the trained model to predict future popularity from current trend data, allowing it to predict which events and content will attract attention in the future. The server adds the prediction results to the trend data and assigns a predicted popularity score to each trend.

[0232] The server recommends advertising

[0233] The server selects trends suitable for advertising based on the predicted popularity, prioritizes trends with high predicted popularity, and recommends the top trends that are expected to be particularly effective as advertising targets.

[0234] Introducing the Emotion Engine

[0235] The server uses an emotion engine to recognize the user's emotions. Specifically, it collects and analyzes the user's emotional data (e.g., comments and reactions posted on social media) to understand the user's emotional state in real time.

[0236] Analyzing Emotional Data

[0237] The server analyzes the collected emotion data and predicts the effectiveness of advertising. For example, it predicts that advertising will be more effective for topics that receive a lot of positive emotion, and prioritizes recommendations for those topics. Conversely, if there are a lot of negative emotions, it decides not to advertise on those topics.

[0238] Advertisement changes and adjustments based on emotions

[0239] The server can also recommend changes to ad content or creative based on user emotion recognition. For example, if a user's emotions are positive, it can suggest ads with positive messages. Conversely, if a user's emotions are predominantly negative, it can suggest ads with more neutral content.

[0240] Specific examples

[0241] For example, the following actions are taken for the 2023 Basketball World Cup:

[0242] 1. The server collects current topics from social media and news APIs:

[0243] The server sends a request to the API to retrieve trending data about the Basketball World Cup. For example, it detects that the hashtag "BasketballWorldCup2023" is popular.

[0244] 2. The server collects and analyzes past event data:

[0245] The server collects data on past sporting events (e.g., the last Basketball World Cup, the Olympics) and preprocesses their viewership and attendance figures.

[0246] 3. The server trains the model:

[0247] The server uses this historical data to train a machine learning model (e.g., a linear regression model).

[0248] 4. The server predicts future popularity:

[0249] The server uses the trained model to predict future popularity based on current trend data, and the prediction reveals that the Basketball World Cup 2023 will peak next weekend.

[0250] 5. The server recommends advertising:

[0251] Based on the prediction results, the server recommends "Basketball World Cup 2023" as an advertising destination. It assigns a score indicating the applicability and priority of the advertising destination and recommends the optimal timing for advertising.

[0252] 6. Introducing the Emotion Engine:

[0253] The server collects user reactions and comments from social media and analyzes them to understand user sentiment. For example, if the reaction to the "Basketball World Cup 2023" is positive, it determines that advertising will be effective.

[0254] 7. Advertisement modification and adjustment based on emotions:

[0255] If the server detects a user's positive emotions, it will suggest ads with positive messages. Conversely, if a user's negative emotions are detected, it will suggest more neutral ads.

[0256] In this way, the system of the present invention can not only respond to rapidly trending content in real time and effectively recommend advertising placements, but also take user emotions into consideration, thereby achieving even more effective advertising operations.

[0257] The processing flow will be explained below.

[0258] Step 1:

[0259] The server sends a request to a social media or news API to retrieve the most popular topics. Specifically, it makes a request to the Twitter API or Google News API to retrieve current trending data. The returned data is received in JSON format.

[0260] Step 2:

[0261] The server parses the API response and converts the trend data from JSON format into Python objects, making it easier to work with in your programs.

[0262] Step 3:

[0263] The server extracts keywords and temporal features (e.g., event dates and times, related hashtags) that are likely to be relevant to advertising from the trend data it acquires. This extraction is performed through text analysis of the trend data.

[0264] Step 4:

[0265] Based on the keywords extracted by the server, past event data is collected from a database or another API, and data on past sporting and entertainment events related to the extracted keywords is obtained.

[0266] Step 5:

[0267] The server preprocesses the collected historical event data. This preprocessing involves extracting features and target variables from data such as viewership, number of participants, and number of posts on social media. This prepares the data in a format suitable for model training.

[0268] Step 6:

[0269] The server uses the preprocessed data to train a machine learning algorithm (e.g., a linear regression model). Through the training process, the model learns the relationship between features and popularity scores.

[0270] Step 7:

[0271] The server uses the trained model to predict future popularity based on current trend data. Specifically, current trend data is input into the model, and predicted viewership ratings and social media reactions are output.

[0272] Step 8:

[0273] The server adds the predicted popularity score to the trend data, which gives each trend a numerical popularity score that can be used to determine advertising placements.

[0274] Step 9:

[0275] The server selects the best trend for advertising based on the popularity score. Specifically, it prioritizes trends with high popularity scores and lists them as recommended advertising destinations.

[0276] Step 10:

[0277] The server obtains user reactions and comments from the social media platform to collect user sentiment data, which includes the process of collecting reaction data through APIs.

[0278] Step 11:

[0279] The server analyzes the collected emotional data to identify the user's emotional state. Sentiment analysis uses natural language processing technology to classify emotions such as positive and negative from the content of comments.

[0280] Step 12:

[0281] The server changes or adjusts advertising destinations based on the emotion data. For example, it determines that topics with a high percentage of positive emotions have a high advertising effectiveness and recommends them as priority destinations.

[0282] Step 13:

[0283] The server combines emotion data with trend data to improve the accuracy of the model, enabling more accurate predictions that take into account the impact of emotional states on predictions.

[0284] Step 14:

[0285] The server notifies the user of recommended advertising placements. Specifically, it presents the user with a list of recommended advertising placements and provides detailed information on the optimal timing and location.

[0286] Step 15:

[0287] After the ad is placed, the server collects actual ad performance data, including viewer ratings, social media responses, and click counts.

[0288] Step 16:

[0289] The server analyzes the performance data collected and analyzes the difference between predicted and actual results, which allows the model to identify areas for improvement.

[0290] Step 17:

[0291] The server uses the feedback to retrain the model and improve its prediction accuracy. By incorporating real data, future predictions will be more accurate.

[0292] Through the above steps, the system of the present invention can respond to rapidly trending content in real time and realize effective advertising that takes into account user emotions.

[0293] Example 2

[0294] 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."

[0295] Conventional advertising systems typically place ads based on past data and current trend data. However, these systems do not take into account user emotional data, which limits the effectiveness of advertising and makes it difficult to optimally place ads. Particularly in today's world, where real-time data such as social media is increasingly being used, ignoring user emotions and reactions will result in a decrease in advertising effectiveness. Therefore, there is a demand for a system that takes user emotions into account and enables more accurate advertising placement.

[0296] 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 a means for collecting past event data, a means for training a model based on the collected past event data, and a means for collecting current trend data. This enables a means for collecting and analyzing user emotion data and a means for adjusting the content and creative of advertisements based on the emotion data.

[0297] "Past event data" refers to information such as viewership ratings, number of participants, number of social media posts, and number of related news articles related to a specific period or event.

[0298] "Training a model" refers to using collected data to train a machine learning algorithm to learn the relationship between features and target variables, making it usable for future predictions.

[0299] "Current trend data" refers to information about the latest topics and trends collected from social media and news APIs.

[0300] "Predicting future popularity" refers to using a trained model to calculate future attention and popularity from current trend data.

[0301] "Recommending advertising destinations" refers to suggesting optimal advertising investment destinations based on predicted popularity.

[0302] "User emotional data" refers to information on emotions and reactions obtained by analyzing comments and reactions posted by users on social media, etc.

[0303] "Collecting and analyzing emotional data" refers to collecting user comments and reactions and analyzing them to understand their emotional state, whether positive, negative, or neutral.

[0304] "Adjusting advertising content and creative based on emotional data" refers to optimizing advertising messages and visuals based on collected and analyzed emotional data.

[0305] This invention combines an advertising recommendation system with an emotion engine that recognizes user emotions. Specifically, it collects past event data, trains a model based on this, collects current trend data to predict future popularity, and recommends advertising destinations. In addition, by utilizing user emotion data, it achieves more accurate advertising placement.

[0306] The server sends requests to social media and news APIs (e.g., Twitter API, NewsAPI) to gather current topics. The data is often returned in JSON format, and the server parses it to extract keywords related to a particular topic or event. For example, the keyword "2023 Basketball World Cup" might be extracted.

[0307] Next, the server uses the keywords extracted from the current trends to collect past event data from a database or another API. This data includes detailed information about the popularity and attention of the event, such as viewership, number of participants, number of social media posts, and number of related news articles. The server preprocesses this data and extracts features and target variables.

[0308] The server then uses the preprocessed historical event data to train a machine learning algorithm (e.g., a linear regression model). Through this process, the model learns the relationship between features and popularity scores. Once trained, the model is used to predict future popularity from current trend data. This prediction process makes it possible to predict which events or content will attract attention in the future. The server adds the prediction results to the trend data and assigns a predicted popularity score to each trend.

[0309] Next, the server selects trends suitable for advertising based on the predicted popularity. For example, it prioritizes popular trends and recommends the top trends that are expected to be particularly effective as advertising targets.

[0310] Furthermore, the server will incorporate an emotion engine to collect and analyze user emotional data. Specifically, it will collect user reactions and comments from social media and analyze them to understand the user's emotional state in real time. For example, it will use IBM Watson's Natural Language Understanding API. This analysis will enable it to predict the effectiveness of advertising, predicting that advertising will be more effective for topics that attract a lot of positive emotions and prioritizing those topics for recommendations. Conversely, if there are a lot of negative emotions, it will decide not to advertise.

[0311] Furthermore, the server can recommend changes to the content and creative of advertisements based on the user's emotional state. For example, if the user's emotional state is positive, the server can suggest advertisements with positive messages. Conversely, if the user's emotional state is mostly negative, the server can suggest advertisements with more neutral content.

[0312] As a concrete example, for the 2023 Basketball World Cup, the following actions are taken:

[0313] The server uses the Twitter API to obtain trend data related to "BasketballWorldCup2023." For example, it detects that the hashtag "BasketballWorldCup2023" is frequently used. Next, the server collects and preprocesses viewership and attendance data related to previous Basketball World Cups and Olympics from Google Analytics. Next, the server trains a linear regression model using this historical data. Using the trained model, it predicts future popularity based on current trend data and predicts that "BasketballWorldCup2023" will peak next weekend. Based on the prediction results, the server then recommends "BasketballWorldCup2023" as an advertising destination. Next, the server uses IBM Watson's API to confirm positive sentiment toward "BasketballWorldCup2023" and suggests ads containing positive messages.

[0314] In this way, the system of the present invention can not only respond to rapidly trending content in real time and effectively recommend advertising placements, but also take user emotions into consideration, thereby achieving even more effective advertising operations.

[0315] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0316] Step 1:

[0317] The server retrieves the current trend.

[0318] The server sends requests to social media and news APIs (e.g., Twitter API, News API) to collect current topics. For example, to obtain trending data about the "Basketball World Cup 2023," the server makes the following API request:

[0319] Input: API request for "BasketballWorldCup2023"

[0320] Output: Trend data in JSON format

[0321] The server parses this JSON data and extracts keywords related to specific topics or events, using Python libraries to parse the JSON data and extract the necessary information.

[0322] Step 2:

[0323] The server collects and analyzes past event data.

[0324] The server uses keywords extracted from current trends and gathers data on past events from a database or another API, for example, data on past Basketball World Cups or Olympic Games.

[0325] Input: Current trending keyword (e.g. "Basketball World Cup")

[0326] Output: Past event data (e.g., viewership, number of participants, number of social media posts)

[0327] The server preprocesses this data and extracts features (e.g., viewership rate, number of SNS posts) and target variables (e.g., popularity score) using Python's pandas library.

[0328] Step 3:

[0329] The server trains the model.

[0330] The server uses the pre-processed historical event data to train a machine learning algorithm (e.g., a linear regression model).

[0331] Input: Preprocessed historical event data (features and target variables)

[0332] Output: The trained model

[0333] Implement and train a linear regression model using Python's scikit-learn library.

[0334] Step 4:

[0335] The server predicts future popularity.

[0336] The server uses the trained model to predict future popularity from current trend data.

[0337] Input: Features of the current trend data

[0338] Output: Predicted popularity score

[0339] The trained model is used to make predictions and assign each trend a predicted popularity score.

[0340] Step 5:

[0341] The server recommends advertising placements.

[0342] The server selects trends suitable for advertising based on the predicted popularity.

[0343] Input: Predicted popularity score

[0344] Output: A list of recommended ad placements

[0345] Based on the prediction results, the most popular trends are prioritized, and the top trends that are expected to be particularly effective are recommended as advertising targets.

[0346] Step 6:

[0347] The server implements the emotion engine.

[0348] The server uses an emotion engine to collect and analyze user emotion data, for example, by collecting user reactions and comments from social media and analyzing their emotions.

[0349] Input: Social media comments and reactions

[0350] Output: Parsed sentiment data (positive, negative, neutral, etc.)

[0351] Sentiment analysis is performed using IBM Watson's Natural Language Understanding API, among others.

[0352] Step 7:

[0353] The server analyzes the emotion data.

[0354] The server analyzes the collected emotional data and predicts the effectiveness of advertising placements.

[0355] Input: Parsed emotion data

[0356] Output: Prediction of advertising effectiveness

[0357] The system predicts that advertising will be effective for topics that have a lot of positive sentiment, and prioritizes recommendations for those topics. If there are a lot of negative sentiments, the system decides not to advertise on those topics.

[0358] Step 8:

[0359] The server modifies and adjusts the ads based on the sentiment.

[0360] The server recommends changes to the content and creative of the advertisement based on the user's emotion recognition.

[0361] Input: Parsed emotion data

[0362] Output: Proposal for changes to ad content

[0363] For example, if you have a lot of positive emotions, we'll suggest ads with positive messages, and if you have a lot of negative emotions, we'll suggest ads with more calming messages.

[0364] (Application example 2)

[0365] 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."

[0366] Previous ad placement recommendation systems focused on predicting ad placements based on past data and current trend data. However, ad placements that do not take user emotions and reactions into account may not achieve the expected results. Furthermore, by taking real-time reactions into account, it is necessary to flexibly adjust the content and timing of ads. Therefore, the challenge is to use user emotion data to achieve more accurate ad placements and maximize advertising effectiveness.

[0367] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting past data, means for training a model based on the collected past data, means for collecting current trend data, means for predicting future popularity from the current trend data using the trained model, means for recommending advertisement placements based on the predicted popularity, means for collecting and analyzing user emotion data, and means for adjusting the content of advertisements based on the analyzed emotion data. This enables highly accurate advertisement placement that takes user emotions into consideration.

[0368] "Past data" refers to data about events or occurrences that occurred in the past, including viewership ratings, number of participants, number of posts on social media, number of related articles, etc.

[0369] "Current trend data" is data about events and content that are currently trending in society and the market, and is obtained from social media and news articles.

[0370] "Future popularity" is a value that predicts how much attention and interest a topic will attract in the future based on current trend data.

[0371] A "model" is a mathematical structure that is trained by a machine learning algorithm to learn patterns from given data and make predictions or classifications.

[0372] "Recommending advertising locations" means proposing the optimal location and timing for placing an ad based on predicted future popularity.

[0373] "User emotional data" refers to data that indicates the emotions and reactions expressed by users on social media or other platforms, including positive and negative emotions.

[0374] An "emotion engine" is software that collects and analyzes a user's emotional data to recognize their emotions.

[0375] "Features" are characteristics or attributes used by models to learn data in data analysis, including things like an event's viewership rating or number of attendees.

[0376] A "target variable" is a variable that is the target of prediction by a prediction model, and in this invention, future popularity is the target variable.

[0377] "Preprocessing" refers to a series of steps taken to convert raw data into an analyzable format, including data cleaning and transformation.

[0378] "Analysis" is the process of extracting useful information and patterns from collected data.

[0379] "Adjusting the content of the advertisement" means changing the message or creative of the advertisement based on the user's emotional data.

[0380] The present invention provides a method for implementing a system that optimizes advertising placement using user emotion data, and an embodiment thereof will be described in detail below.

[0381] The server collects current trend data from social media and news APIs. For example, it retrieves how often a particular hashtag is used or which topics are mentioned in news articles. This data is returned in JSON format and analyzed. Key keywords and temporal features are extracted during analysis. This process is performed using Python's requests library and data analysis libraries (e.g., Pandas).

[0382] Next, the server collects past event data, including viewership, number of participants, number of social media posts, number of related news articles, etc. The collected data is preprocessed to extract features and target variables. This is done using a machine learning library (e.g., scikit-learn).

[0383] The server trains a machine learning model based on the preprocessed historical data. The training process uses the features as input and outputs a popularity score. The trained model is then applied to a linear regression model or other regression models.

[0384] The server also uses current trend data to predict future popularity, allowing it to predict which events and content will attract attention in the future. The predicted popularity is output in JSON format, and a score is assigned to each trend.

[0385] Furthermore, the server collects and analyzes users' emotional data. A sentiment analysis engine is used to analyze users' social media posts and reactions. This engine typically uses natural language processing libraries (e.g., Hugging Face transformers).

[0386] The server applies an algorithm to adjust the content of ads based on the emotional data. For example, ads with positive messages are delivered to users based on positive emotional data, while ads with more neutral content are recommended for users based on negative emotional data. Existing ad management systems (e.g., Google Ads) can be used to adjust ads.

[0387] A concrete example of server processing is the 2023 Basketball World Cup. The server collects current topics from social media and news APIs and detects that the hashtag "BasketballWorldCup2023" is frequently used. It collects data on past sporting events and preprocesses their viewership and attendance figures. It then uses a machine learning model to predict future popularity and predicts that "BasketballWorldCup2023" will peak next weekend. It also collects sentiment data from users' social media posts, and if there are a lot of positive reactions, it delivers ads with positive messages, thereby increasing advertising effectiveness.

[0388] Examples of prompts include:

[0389] Collect current trend data and use it to predict future popularity. Additionally, collect user reactions from social media and analyze sentiment. Based on the results, recommend the best advertising destinations.

[0390] As described above, the system of the present invention provides a technique for utilizing user emotion data to realize effective and highly accurate advertising.

[0391] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0392] Step 1:

[0393] The server collects current trend data from social media and news APIs. Specifically, it sends requests to the APIs and obtains the returned JSON-formatted data. The input is the endpoints of the social media and news APIs, and the output is JSON-formatted data about current topics. A data analysis library (e.g., Pandas) is used to extract key keywords and temporal features.

[0394] Step 2:

[0395] The server collects past event data. It retrieves event viewership, number of participants, number of social media posts, number of related news articles, etc. from a database or another API. The input is queries or API requests related to past events, and the output is past event data. This data is preprocessed to extract features and target variables. Preprocessing is performed using a machine learning library (e.g., scikit-learn).

[0396] Step 3:

[0397] The server trains a machine learning model based on preprocessed historical data. The input is the preprocessed features and target variable, and the output is the trained model. The model can be a linear regression model or another regression model. A machine learning library (e.g., scikit-learn) is used.

[0398] Step 4:

[0399] The server uses current trend data to predict future popularity. The input is the current trend data and a trained model, and the output is the future popularity score for each trend. The preprocessed current trend data is input into the machine learning model to perform predictions. The prediction results are output in JSON format.

[0400] Step 5:

[0401] The server collects and analyzes user emotion data. The input is user posts and reaction data from social media, and the output is the analyzed emotion data. An emotion analysis engine (e.g., Hugging Face transformers) is used to recognize positive, negative, and neutral emotions in real time.

[0402] Step 6:

[0403] The server recommends ad placements based on future popularity scores and user emotion data. The input is future popularity scores and analyzed emotion data, and the output is a list of optimal ad placements. Connected to the ad management system, the recommendation engine suggests optimal ad placement locations and timing.

[0404] Step 7:

[0405] The server adjusts the content of the advertisements based on the emotional data. The input is the analyzed emotional data, and the output is the adjusted advertising content. If the user's emotional data is positive, advertisements with positive messages are delivered, and if the emotional data is negative, advertisements with gentle content are recommended. This process uses an existing advertising management system (e.g., Google Ads Management).

[0406] The above are the specific processing steps for carrying out the present invention.

[0407] 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.

[0408] 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.

[0409] 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.

[0410] [Second embodiment]

[0411] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0412] 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.

[0413] 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).

[0414] 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.

[0415] 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.

[0416] 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).

[0417] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for 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.

[0418] 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.

[0419] 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.

[0420] 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.

[0421] In the smart glasses 214, the 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.

[0422] 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."

[0423] The present invention relates to a system for recommending advertisement placements, and includes processes for collecting past event data, training a model, collecting current trend data, predicting future popularity based on that data, and recommending advertisement placements. The program processing of this system is explained below in natural language.

[0424] Program Overview

[0425] The server retrieves the current trend

[0426] The server sends requests to social media and news APIs to retrieve current topics. This data, which contains information about specific topics or events, is returned in JSON format. The server analyzes this trend data to extract key keywords and temporal features.

[0427] The server collects and analyzes past event data

[0428] The server uses keywords extracted from current trends to collect past event data from a database or another API. This past data includes information about the popularity and attention of the event, such as viewer ratings, number of participants, number of social media posts, and number of related news articles. The server preprocesses this data and extracts features and target variables.

[0429] The server trains the model

[0430] The server uses the preprocessed historical event data to train a machine learning algorithm (e.g., a linear regression model). During this training process, features are input, popularity scores are output, and the model learns the relationship between them.

[0431] The server predicts future popularity

[0432] The server uses the trained model to predict future popularity from current trend data, allowing it to predict which events and content will attract attention in the future. The server adds the prediction results to the trend data and assigns a predicted popularity score to each trend.

[0433] The server recommends advertising

[0434] The server selects trends suitable for advertising based on the predicted popularity, prioritizes trends with high predicted popularity, and recommends the top trends that are expected to be particularly effective as advertising targets.

[0435] Specific examples

[0436] For example, if there is a sudden surge in excitement for the 2023 Basketball World Cup, the system would work as follows:

[0437] 1. The server collects current topics about the Basketball World Cup from social media and news APIs. For example, it detects that the hashtag "BasketballWorldCup2023" is frequently used.

[0438] 2. The server collects past sporting event data (e.g., the last Basketball World Cup, the Olympics) and preprocesses their viewership and attendance figures.

[0439] 3. The server uses this historical data to train a machine learning model (e.g., a linear regression model).

[0440] 4. The server uses the trained model to predict future popularity based on current trend data. The prediction reveals that the Basketball World Cup 2023 will peak next weekend.

[0441] 5. Based on the prediction results, the server recommends "Basketball World Cup 2023" as an advertising destination. It assigns a score indicating the applicability and priority of the advertising destination and recommends the optimal timing for advertising.

[0442] 6. After the ads are actually placed, the server collects the results and feeds them back into the model to improve prediction accuracy.

[0443] In this way, the system of the present invention can respond to trending content in real time and effectively recommend advertising, allowing advertisers to maximize their business opportunities without missing out.

[0444] The processing flow will be explained below.

[0445] Program processing steps

[0446] Step 1:

[0447] The server sends a request to a social networking or news API, such as the Twitter API or Google News API, to get the most popular topics at the moment, which allows you to get current trending data.

[0448] Step 2:

[0449] The server receives the API response and retrieves the trending data in JSON format, including specific keywords, topics, hashtags, etc.

[0450] Step 3:

[0451] The server analyzes the trend data and extracts keywords and temporal features (e.g., event dates and times, related hashtags) that are likely to be relevant to advertising. This data is used for subsequent model training.

[0452] Step 4:

[0453] The server then collects past event data from a database or other APIs based on the extracted keywords. For example, trending keywords related to the Basketball World Cup are used to collect data on similar past sporting events.

[0454] Step 5:

[0455] The server preprocesses past event data collected by the server. Specifically, it extracts features and target variables from data such as viewership, number of participants, and number of posts on social media. This creates a dataset in a format suitable for learning.

[0456] Step 6:

[0457] The server uses the preprocessed data to train a machine learning algorithm (e.g., a linear regression model). Through this training process, the model learns the relationship between features and popularity scores.

[0458] Step 7:

[0459] The server uses the trained model to predict future popularity based on current trend data. For example, it uses the current trend data of the Basketball World Cup as input and predicts future viewership and the likelihood of it becoming a hot topic on social media.

[0460] Step 8:

[0461] The server adds the predictions to the trend data and gives each trend a predicted popularity score, which reveals which trends are most suitable for advertising.

[0462] Step 9:

[0463] The server selects trends suitable for advertising based on predicted popularity, prioritizes trends with high predicted popularity, and recommends the top trends that are expected to be particularly effective as advertising targets.

[0464] Step 10:

[0465] The server notifies the user of the ad placement destinations. Specifically, it presents the user with a list of recommended ad placement destinations, along with details of the timing and location.

[0466] Step 11:

[0467] After the ad is placed, the server collects actual popularity data, such as viewer ratings and social media reactions after the ad is placed.

[0468] Step 12:

[0469] The server analyzes the difference between the actual popularity data and the predictions to identify areas for improvement in the model, providing feedback to improve the accuracy of the prediction model.

[0470] Step 13:

[0471] The server retrains the model using actual popularity data to improve prediction accuracy, enabling more accurate predictions for future ad placements.

[0472] Through the steps described above, the system of the present invention can respond to rapidly trending content in real time and effectively recommend advertising placements.

[0473] Example 1

[0474] 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."

[0475] Conventional advertising systems do not fully utilize past data and lack real-time trend analysis. As a result, they are unable to optimally select advertising destinations, resulting in reduced advertising effectiveness. Furthermore, they lack a feedback function to maintain and improve the accuracy of their predictive models, making it difficult to maximize advertising effectiveness over the long term.

[0476] 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.

[0477] In this invention, the server includes means for collecting past data, means for training a model based on the collected past data, means for collecting current trend data, means for predicting future popularity from the current trend data using the trained model, means for recommending advertising destinations based on the predicted popularity, means for acquiring and analyzing trend data from an API, means for preprocessing past data and extracting features and target variables, means for assigning popularity scores to each trend, and means for collecting actual results of advertising as feedback and improving the model. This enables advertisers to respond to rapidly trending content in real time and place effective advertisements.

[0478] "Past data" refers to information about past events, including viewership ratings, number of participants, number of posts on social media, and number of related news articles.

[0479] "Training a model" means using collected data to build a computational model using machine learning algorithms to learn patterns from the data and make predictions.

[0480] "Current trend data" is real-time information related to specific topics or events obtained from social media and news sites.

[0481] "Predicting future popularity" means using a trained model to predict the popularity of future events or topics based on current trend data.

[0482] "Recommending advertising destinations" means suggesting which events or topics you should advertise on based on predicted future popularity information.

[0483] "Obtaining and analyzing trend data from an API" means using an application programming interface (API) to obtain data from social media and news sites, and analyzing that data to extract useful information.

[0484] A "feature" is a specific attribute or characteristic of data that is used as input for a machine learning model.

[0485] A "target variable" is a value or category that is predicted as the output of a machine learning model.

[0486] A "Popularity Score" is a numerical representation of how much attention a particular event or topic will receive in the future.

[0487] "Feedback" refers to data collected from the results of actual advertising placements (e.g., click-through rate, conversion rate) and used to improve the performance of the model.

[0488] The present invention relates to a system for recommending advertisements, which trains a machine learning model based on past data and current trend data, predicts future popularity, and recommends advertisement placements. Specific embodiments for implementing this system are described below.

[0489] System configuration and technologies used

[0490] The server sends requests to the APIs of social media platforms and news services (e.g., Twitter API, Google News API) to obtain current trending data, which is returned in JSON format and parsed by the server to extract information about specific topics or events.

[0491] The server uses the extracted keywords to collect past data from a database or another API (e.g., sports database API, Google Trends API). The collected data includes viewership, number of participants, number of social media posts, number of related news articles, etc. The server preprocesses this data and extracts features and target variables. Specifically, it imputes missing values ​​and normalizes the data.

[0492] The server uses the preprocessed data to train a model using a machine learning algorithm (e.g., linear regression model or random forest). This training process uses the Scikit-learn library, with features as input and popularity scores as output. The dataset is split into training data and test data, and the performance of the model is evaluated.

[0493] The server uses the trained model to predict future popularity based on current trend data. Specifically, it inputs data from the "Basketball World Cup 2023" obtained from the Twitter API into the model and calculates the predicted level of attention for the following weekend. The output is a popularity score, which is used to evaluate future trends.

[0494] The server recommends optimal advertising destinations based on the predicted popularity score. For example, if interest in the "Basketball World Cup 2023" is predicted to be at its highest next weekend, the server will recommend this trend as an advertising destination. It also provides specific advertising strategies and timing to advertisers.

[0495] The server collects the results data (e.g., click-through rate, conversion rate) after the ads are actually placed and uses them as feedback to improve the accuracy of the model. The results data are used to retrain the model, maximizing the effectiveness of advertising in the long term.

[0496] Examples of specific examples and prompts

[0497] For example, if the 2023 Basketball World Cup is trending rapidly on social media, the system would work like this:

[0498] 1. The server collects current trend data related to "BasketballWorldCup2023" from the Twitter API and Google News API, and checks the frequency of the extracted keyword "BasketballWorldCup2023."

[0499] 2. The server collects viewership data and participant numbers from the previous Basketball World Cup from the sports database API, completes the missing data, and extracts it as features.

[0500] 3. The server uses Scikit-learn to train a linear regression model based on past data and build a model to predict the popularity of events.

[0501] 4. The server uses the trained model to predict future interest in the Basketball World Cup 2023 and estimates the peak interest for the following weekend.

[0502] 5. Based on the prediction results, the server recommends "Basketball World Cup 2023" as an advertising destination and proposes specific advertising timing and strategies to the advertiser.

[0503] 6. The server collects post-advertising result data (click-through rate and conversion rate) and retrains the predictive model to improve accuracy.

[0504] Example prompt sentence:

[0505] "Collect current social media trends related to the 2023 Basketball World Cup, use past sporting event data to predict future popularity, and recommend optimal advertising destinations. Specifically, retrieve trend data from the Twitter API, train a linear regression model using Scikit-learn, and propose an advertising strategy based on the prediction results."

[0506] In this way, the system of the present invention provides advertisers with a means to effectively place advertisements in response to rapidly trending content in real time.

[0507] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0508] Step 1:

[0509] The server collects current trend data. It sends requests to APIs of social media platforms and news services (e.g., Twitter API, Google News API) to retrieve data about current topics. The input is the API request, and the output is trend data in JSON format. The server parses this data to extract key information about specific topics or events, such as the frequency of occurrence of the hashtag "BasketballWorldCup2023."

[0510] Step 2:

[0511] The server collects past data. Using the keywords extracted in step 1, the server retrieves past related event data from a database or another API (e.g., sports database API, Google Trends API). This includes viewership, number of participants, number of social media posts, number of related news articles, etc. The input is keywords related to past events, and the output is a dataset of related past events. The server parses this dataset to extract the necessary information.

[0512] Step 3:

[0513] The server preprocesses the data. The server normalizes the past data collected in step 2, fills in missing values, and formats it as features and target variables. For example, it formats viewer ratings and number of posts as numerical data. The input is raw data of past events, and the output is a preprocessed dataset.

[0514] Step 4:

[0515] The server trains the model. The server uses the preprocessed data to train the model using a machine learning algorithm (e.g., linear regression model or random forest). For example, it uses the Scikit-learn library. The input is the preprocessed dataset, and the output is the trained predictive model. The model is configured to input features and output a popularity score.

[0516] Step 5:

[0517] The server predicts future popularity. The server uses the trained model to predict future popularity from the current trend data obtained in step 1. The input is the current trend data and the trained model, and the output is the future popularity score. For example, the server predicts that the hashtag "BasketballWorldCup2023" will peak next weekend.

[0518] Step 6:

[0519] The server recommends advertising destinations. The server selects the optimal advertising destination based on the popularity score predicted in step 5 and recommends advertising to the advertiser. The input is the future popularity score, and the output is the recommended advertising destination. Specifically, the server recommends "Basketball World Cup 2023" as an advertising destination and suggests advertising strategies and timing.

[0520] Step 7:

[0521] The server collects feedback and improves the model. The server collects results after the ads are actually placed (e.g., click-through rate, conversion rate), and uses that data to retrain the model and improve its accuracy. The input is the results data after the ads are placed, and the output is an improved model. This makes it possible to maximize advertising effectiveness in the long term.

[0522] (Application example 1)

[0523] 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."

[0524] Conventional advertising systems statically select targets based on past data and are unable to respond to real-time trend changes, making it difficult to optimize advertising effectiveness. Furthermore, there is no feedback based on post-advertising results to improve model accuracy, meaning long-term improvements in advertising effectiveness cannot be expected. Furthermore, there is a lack of a way for users to easily evaluate advertising effectiveness and reflect that in their next ad placement.

[0525] 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.

[0526] In this invention, the server includes means for collecting past event data, means for training a model based on the collected past event data, means for collecting current trend data, means for predicting future popularity from the current trend data using the trained model, means for recommending advertising destinations based on the predicted popularity, means for users to provide feedback on the results of their advertising placements, and means for improving the accuracy of the model using the feedback results. This enables rapid response to trend fluctuations in real time and optimization of advertising effectiveness. Furthermore, improving the accuracy of the model based on user feedback can also be expected to improve long-term advertising effectiveness. Furthermore, users can easily evaluate advertising effectiveness and reflect it in their next advertising placements.

[0527] "Past event data" refers to information related to events that have taken place in the past, including viewership ratings, number of participants, number of posts on social media, number of related news articles, and the like.

[0528] "Training a model" is the process of using machine learning algorithms to learn relationships between collected data.

[0529] "Current trending data" is data on information and topics that are currently trending, collected from social media, news, and other sources.

[0530] "Predicting future popularity" is the process of using a trained model to infer what will be popular and trending in the future based on current trend data.

[0531] "Recommending advertising locations" means suggesting the most effective locations and timing for advertising based on predicted popularity.

[0532] "Providing feedback" means that the user reports to the server the results and effects obtained after placing an advertisement.

[0533] "Improving the accuracy of the model" means adjusting the parameters and algorithms of the predictive model based on the feedback results to increase the accuracy of the next prediction.

[0534] This invention relates to a system for recommending advertising, which consists of the following steps: First, a server collects past event data and trains a machine learning model based on this data. Next, the server collects current trend data and uses the trained model to predict future popularity. Based on the prediction results, the system recommends optimal advertising destinations to users.

[0535] Program processing

[0536] Data Collection Module

[0537] The server collects current trend data via social media and news APIs. Specifically, it uses Python to send requests to the Twitter API and Google News API, and retrieves JSON-formatted data including trending keywords and the number of posts.

[0538] Hardware used: Server, smartphone

[0539] Software used: Python, Twitter API, Google News API

[0540] Historical Data Analysis Module

[0541] The server collects past event data from a database or another API, and performs preprocessing to extract features and target variables. In this process, metrics such as viewership, number of participants, and number of posts on social media are used, and the data is organized using Python's Scikit-learn.

[0542] Hardware used: Server

[0543] Software used: Python, Scikit-learn

[0544] Model Training Module

[0545] The server uses the preprocessed data to train a machine learning algorithm (e.g., a linear regression model), which learns relationships in the data and generates a model to predict future popularity.

[0546] Hardware used: Server

[0547] Software used: Python, Scikit-learn

[0548] Popularity Prediction Module

[0549] The server uses the trained model to predict future popularity based on current trend data, a process also performed using Python.

[0550] Hardware used: Server

[0551] Software used: Python, Scikit-learn

[0552] Recommendation Engine

[0553] The server recommends the most effective advertising destinations based on predicted popularity, allowing users to know effective advertising destinations in real time.

[0554] Hardware used: Server

[0555] Software used: Python, Flask

[0556] Feedback Module

[0557] Users provide feedback on the results of their ad placements, and the data is sent to the server, where it is used to retrain the model, improving its prediction accuracy.

[0558] Hardware used: Smartphone

[0559] Software used: React Native, Firebase

[0560] Specific examples

[0561] As a specific application example, we will explain the process in which a server collects the latest trend data and predicts future popularity. For example, the server inputs the following prompt sentence into the generative AI model:

[0562] Gather the latest trends and predict future popularity. Use past event data to identify hot topics for the upcoming weekend and suggest them for advertising.

[0563] Through this process, the server can quickly respond to trend fluctuations in real time and optimize advertising effectiveness. It also has a feedback function to improve the accuracy of the model, making it a system that can achieve long-term improvements in advertising effectiveness.

[0564] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0565] Step 1:

[0566] The server collects past event data. Specifically, the server retrieves data related to past events, such as viewer ratings, number of participants, and number of posts on social media, from a database or API. After collection, this data is saved in JSON format and used for subsequent processing.

[0567] Input: Historical event data from a database or API

[0568] Output: Historical event data stored in JSON format

[0569] Step 2:

[0570] The server preprocesses the collected past event data and extracts features and target variables. Specifically, it performs missing value imputation, noise removal, and scaling of the data, and then extracts features (e.g., viewership rate, number of participants, number of social media posts) and target variables. Python's Scikit-learn is used.

[0571] Input: Historical event data stored in JSON format

[0572] Output: Preprocessed features and target variable

[0573] Step 3:

[0574] The server trains the model using the preprocessed data. Specifically, it uses a machine learning algorithm such as a linear regression model, inputs features, and outputs target variables. Python's Scikit-learn is used.

[0575] Input: Preprocessed features and target variables

[0576] Output: A trained machine learning model

[0577] Step 4:

[0578] The server collects current trend data from social media and news APIs. Specifically, it sends an API request and receives data in JSON format containing the latest topics and keywords. For example, it uses the Twitter API or Google News API.

[0579] Input: Trend data from Twitter API and Google News API

[0580] Output: Current trend data saved in JSON format

[0581] Step 5:

[0582] The server preprocesses the collected current trend data and extracts features. Specifically, it performs noise removal and normalization, and extracts features (e.g., topics, keywords, and number of posts). It uses Python's Scikit-learn.

[0583] Input: Current trend data stored in JSON format

[0584] Output: Preprocessed features

[0585] Step 6:

[0586] The server uses the trained model to predict future popularity based on current trend data. Specifically, the preprocessed features are input into the model to predict future popularity scores.

[0587] Input: Preprocessed features, trained model

[0588] Output: Predicted future popularity scores

[0589] Step 7:

[0590] The server recommends advertising destinations based on the predicted popularity. Specifically, it prioritizes trends with high popularity scores and suggests them as advertising destinations. Users receive this information through a smartphone application.

[0591] Input: Predicted future popularity scores

[0592] Output: Recommended ad placements

[0593] Step 8:

[0594] Users provide feedback on the results of their ads after placing them. Specifically, they use their smartphones to input advertising effectiveness measurement data (click rates, conversion rates, etc.) and send it to the server.

[0595] Input: Post-advertising effectiveness measurement data

[0596] Output: Feedback data sent to the server

[0597] Step 9:

[0598] The server uses the feedback data to improve the accuracy of the model. Specifically, it uses the feedback data as retraining data to adjust the model parameters and algorithms, thereby improving the accuracy of the next prediction.

[0599] Input: Feedback data

[0600] Output: A trained model with improved accuracy

[0601] 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.

[0602] The present invention combines an advertising recommendation system with an emotion engine that recognizes user emotions, and in addition to the process of collecting past event data, training a model based on this, collecting current trend data to predict future popularity, and recommending advertising destinations, the system utilizes user emotion data to achieve more accurate advertising placement. The program processing of this system is explained below in natural language.

[0603] Program Overview

[0604] The server retrieves the current trend

[0605] The server sends requests to social media and news APIs to retrieve current topics. This data, which contains information about specific topics or events, is returned in JSON format. The server analyzes this trend data to extract key keywords and temporal features.

[0606] The server collects and analyzes past event data

[0607] The server uses keywords extracted from current trends to collect past event data from a database or another API. This past data includes information about the popularity and attention of the event, such as viewer ratings, number of participants, number of social media posts, and number of related news articles. The server preprocesses this data and extracts features and target variables.

[0608] The server trains the model

[0609] The server uses the preprocessed historical event data to train a machine learning algorithm (e.g., a linear regression model). During this training process, features are input, popularity scores are output, and the model learns the relationship between them.

[0610] The server predicts future popularity

[0611] The server uses the trained model to predict future popularity from current trend data, allowing it to predict which events and content will attract attention in the future. The server adds the prediction results to the trend data and assigns a predicted popularity score to each trend.

[0612] The server recommends advertising

[0613] The server selects trends suitable for advertising based on the predicted popularity, prioritizes trends with high predicted popularity, and recommends the top trends that are expected to be particularly effective as advertising targets.

[0614] Introducing the Emotion Engine

[0615] The server uses an emotion engine to recognize the user's emotions. Specifically, it collects and analyzes the user's emotional data (e.g., comments and reactions posted on social media) to understand the user's emotional state in real time.

[0616] Analyzing Emotional Data

[0617] The server analyzes the collected emotion data and predicts the effectiveness of advertising. For example, it predicts that advertising will be more effective for topics that receive a lot of positive emotion, and prioritizes recommendations for those topics. Conversely, if there are a lot of negative emotions, it decides not to advertise on those topics.

[0618] Advertisement changes and adjustments based on emotions

[0619] The server can also recommend changes to ad content or creative based on user emotion recognition. For example, if a user's emotions are positive, it can suggest ads with positive messages. Conversely, if a user's emotions are predominantly negative, it can suggest ads with more neutral content.

[0620] Specific examples

[0621] For example, the following actions are taken for the 2023 Basketball World Cup:

[0622] 1. The server collects current topics from social media and news APIs:

[0623] The server sends a request to the API to retrieve trending data about the Basketball World Cup. For example, it detects that the hashtag "BasketballWorldCup2023" is popular.

[0624] 2. The server collects and analyzes past event data:

[0625] The server collects data on past sporting events (e.g., the last Basketball World Cup, the Olympics) and preprocesses their viewership and attendance figures.

[0626] 3. The server trains the model:

[0627] The server uses this historical data to train a machine learning model (e.g., a linear regression model).

[0628] 4. The server predicts future popularity:

[0629] The server uses the trained model to predict future popularity based on current trend data, and the prediction reveals that the Basketball World Cup 2023 will peak next weekend.

[0630] 5. The server recommends advertising:

[0631] Based on the prediction results, the server recommends "Basketball World Cup 2023" as an advertising destination. It assigns a score indicating the applicability and priority of the advertising destination and recommends the optimal timing for advertising.

[0632] 6. Introducing the Emotion Engine:

[0633] The server collects user reactions and comments from social media and analyzes them to understand user sentiment. For example, if the reaction to the "Basketball World Cup 2023" is positive, it determines that advertising will be effective.

[0634] 7. Advertisement modification and adjustment based on emotions:

[0635] If the server detects a user's positive emotions, it will suggest ads with positive messages. Conversely, if a user's negative emotions are detected, it will suggest more neutral ads.

[0636] In this way, the system of the present invention can not only respond to rapidly trending content in real time and effectively recommend advertising placements, but also take user emotions into consideration, thereby achieving even more effective advertising operations.

[0637] The processing flow will be explained below.

[0638] Step 1:

[0639] The server sends a request to a social media or news API to retrieve the most popular topics. Specifically, it makes a request to the Twitter API or Google News API to retrieve current trending data. The returned data is received in JSON format.

[0640] Step 2:

[0641] The server parses the API response and converts the trend data from JSON format into Python objects, making it easier to work with in your programs.

[0642] Step 3:

[0643] The server extracts keywords and temporal features (e.g., event dates and times, related hashtags) that are likely to be relevant to advertising from the trend data it acquires. This extraction is performed through text analysis of the trend data.

[0644] Step 4:

[0645] Based on the keywords extracted by the server, past event data is collected from a database or another API, and data on past sporting and entertainment events related to the extracted keywords is obtained.

[0646] Step 5:

[0647] The server preprocesses the collected historical event data. This preprocessing involves extracting features and target variables from data such as viewership, number of participants, and number of posts on social media. This prepares the data in a format suitable for model training.

[0648] Step 6:

[0649] The server uses the preprocessed data to train a machine learning algorithm (e.g., a linear regression model). Through the training process, the model learns the relationship between features and popularity scores.

[0650] Step 7:

[0651] The server uses the trained model to predict future popularity based on current trend data. Specifically, current trend data is input into the model, and predicted viewership ratings and social media reactions are output.

[0652] Step 8:

[0653] The server adds the predicted popularity score to the trend data, which gives each trend a numerical popularity score that can be used to determine advertising placements.

[0654] Step 9:

[0655] The server selects the best trend for advertising based on the popularity score. Specifically, it prioritizes trends with high popularity scores and lists them as recommended advertising destinations.

[0656] Step 10:

[0657] The server obtains user reactions and comments from the social media platform to collect user sentiment data, which includes the process of collecting reaction data through APIs.

[0658] Step 11:

[0659] The server analyzes the collected emotional data to identify the user's emotional state. Sentiment analysis uses natural language processing technology to classify emotions such as positive and negative from the content of comments.

[0660] Step 12:

[0661] The server changes or adjusts advertising destinations based on the emotion data. For example, it determines that topics with a high percentage of positive emotions have a high advertising effectiveness and recommends them as priority destinations.

[0662] Step 13:

[0663] The server combines emotion data with trend data to improve the accuracy of the model, enabling more accurate predictions that take into account the impact of emotional states on predictions.

[0664] Step 14:

[0665] The server notifies the user of recommended advertising placements. Specifically, it presents the user with a list of recommended advertising placements and provides detailed information on the optimal timing and location.

[0666] Step 15:

[0667] After the ad is placed, the server collects actual ad performance data, including viewer ratings, social media responses, and click counts.

[0668] Step 16:

[0669] The server analyzes the performance data collected and analyzes the difference between predicted and actual results, which allows the model to identify areas for improvement.

[0670] Step 17:

[0671] The server uses the feedback to retrain the model and improve its prediction accuracy. By incorporating real data, future predictions will be more accurate.

[0672] Through the above steps, the system of the present invention can respond to rapidly trending content in real time and realize effective advertising that takes into account user emotions.

[0673] Example 2

[0674] 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."

[0675] Conventional advertising systems typically place ads based on past data and current trend data. However, these systems do not take into account user emotional data, which limits the effectiveness of advertising and makes it difficult to optimally place ads. Particularly in today's world, where real-time data such as social media is increasingly being used, ignoring user emotions and reactions will result in a decrease in advertising effectiveness. Therefore, there is a demand for a system that takes user emotions into account and enables more accurate advertising placement.

[0676] 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 a means for collecting past event data, a means for training a model based on the collected past event data, and a means for collecting current trend data. This enables a means for collecting and analyzing user emotion data and a means for adjusting the content and creative of advertisements based on the emotion data.

[0677] "Past event data" refers to information such as viewership ratings, number of participants, number of social media posts, and number of related news articles related to a specific period or event.

[0678] "Training a model" refers to using collected data to train a machine learning algorithm to learn the relationship between features and target variables, making it usable for future predictions.

[0679] "Current trend data" refers to information about the latest topics and trends collected from social media and news APIs.

[0680] "Predicting future popularity" refers to using a trained model to calculate future attention and popularity from current trend data.

[0681] "Recommending advertising destinations" refers to suggesting optimal advertising investment destinations based on predicted popularity.

[0682] "User emotional data" refers to information on emotions and reactions obtained by analyzing comments and reactions posted by users on social media, etc.

[0683] "Collecting and analyzing emotional data" refers to collecting user comments and reactions and analyzing them to understand their emotional state, whether positive, negative, or neutral.

[0684] "Adjusting advertising content and creative based on emotional data" refers to optimizing advertising messages and visuals based on collected and analyzed emotional data.

[0685] This invention combines an advertising recommendation system with an emotion engine that recognizes user emotions. Specifically, it collects past event data, trains a model based on this, collects current trend data to predict future popularity, and recommends advertising destinations. In addition, by utilizing user emotion data, it achieves more accurate advertising placement.

[0686] The server sends requests to social media and news APIs (e.g., Twitter API, NewsAPI) to gather current topics. The data is often returned in JSON format, and the server parses it to extract keywords related to a particular topic or event. For example, the keyword "2023 Basketball World Cup" might be extracted.

[0687] Next, the server uses the keywords extracted from the current trends to collect past event data from a database or another API. This data includes detailed information about the popularity and attention of the event, such as viewership, number of participants, number of social media posts, and number of related news articles. The server preprocesses this data and extracts features and target variables.

[0688] The server then uses the preprocessed historical event data to train a machine learning algorithm (e.g., a linear regression model). Through this process, the model learns the relationship between features and popularity scores. Once trained, the model is used to predict future popularity from current trend data. This prediction process makes it possible to predict which events or content will attract attention in the future. The server adds the prediction results to the trend data and assigns a predicted popularity score to each trend.

[0689] Next, the server selects trends suitable for advertising based on the predicted popularity. For example, it prioritizes popular trends and recommends the top trends that are expected to be particularly effective as advertising targets.

[0690] Furthermore, the server will incorporate an emotion engine to collect and analyze user emotional data. Specifically, it will collect user reactions and comments from social media and analyze them to understand the user's emotional state in real time. For example, it will use IBM Watson's Natural Language Understanding API. This analysis will enable it to predict the effectiveness of advertising, predicting that advertising will be more effective for topics that attract a lot of positive emotions and prioritizing those topics for recommendations. Conversely, if there are a lot of negative emotions, it will decide not to advertise.

[0691] Furthermore, the server can recommend changes to the content and creative of advertisements based on the user's emotional state. For example, if the user's emotional state is positive, the server can suggest advertisements with positive messages. Conversely, if the user's emotional state is mostly negative, the server can suggest advertisements with more neutral content.

[0692] As a concrete example, for the 2023 Basketball World Cup, the following actions are taken:

[0693] The server uses the Twitter API to obtain trend data related to "BasketballWorldCup2023." For example, it detects that the hashtag "BasketballWorldCup2023" is frequently used. Next, the server collects and preprocesses viewership and attendance data related to previous Basketball World Cups and Olympics from Google Analytics. Next, the server trains a linear regression model using this historical data. Using the trained model, it predicts future popularity based on current trend data and predicts that "BasketballWorldCup2023" will peak next weekend. Based on the prediction results, the server then recommends "BasketballWorldCup2023" as an advertising destination. Next, the server uses IBM Watson's API to confirm positive sentiment toward "BasketballWorldCup2023" and suggests ads containing positive messages.

[0694] In this way, the system of the present invention can not only respond to rapidly trending content in real time and effectively recommend advertising placements, but also take user emotions into consideration, thereby achieving even more effective advertising operations.

[0695] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0696] Step 1:

[0697] The server retrieves the current trend.

[0698] The server sends requests to social media and news APIs (e.g., Twitter API, News API) to collect current topics. For example, to obtain trending data about the "Basketball World Cup 2023," the server makes the following API request:

[0699] Input: API request for "BasketballWorldCup2023"

[0700] Output: Trend data in JSON format

[0701] The server parses this JSON data and extracts keywords related to specific topics or events, using Python libraries to parse the JSON data and extract the necessary information.

[0702] Step 2:

[0703] The server collects and analyzes past event data.

[0704] The server uses keywords extracted from current trends and gathers data on past events from a database or another API, for example, data on past Basketball World Cups or Olympic Games.

[0705] Input: Current trending keyword (e.g. "Basketball World Cup")

[0706] Output: Past event data (e.g., viewership, number of participants, number of social media posts)

[0707] The server preprocesses this data and extracts features (e.g., viewership rate, number of SNS posts) and target variables (e.g., popularity score) using Python's pandas library.

[0708] Step 3:

[0709] The server trains the model.

[0710] The server uses the pre-processed historical event data to train a machine learning algorithm (e.g., a linear regression model).

[0711] Input: Preprocessed historical event data (features and target variables)

[0712] Output: The trained model

[0713] Implement and train a linear regression model using Python's scikit-learn library.

[0714] Step 4:

[0715] The server predicts future popularity.

[0716] The server uses the trained model to predict future popularity from current trend data.

[0717] Input: Features of the current trend data

[0718] Output: Predicted popularity score

[0719] The trained model is used to make predictions and assign each trend a predicted popularity score.

[0720] Step 5:

[0721] The server recommends advertising placements.

[0722] The server selects trends suitable for advertising based on the predicted popularity.

[0723] Input: Predicted popularity score

[0724] Output: A list of recommended ad placements

[0725] Based on the prediction results, the most popular trends are prioritized, and the top trends that are expected to be particularly effective are recommended as advertising targets.

[0726] Step 6:

[0727] The server implements the emotion engine.

[0728] The server uses an emotion engine to collect and analyze user emotion data, for example, by collecting user reactions and comments from social media and analyzing their emotions.

[0729] Input: Social media comments and reactions

[0730] Output: Parsed sentiment data (positive, negative, neutral, etc.)

[0731] Sentiment analysis is performed using IBM Watson's Natural Language Understanding API, among others.

[0732] Step 7:

[0733] The server analyzes the emotion data.

[0734] The server analyzes the collected emotional data and predicts the effectiveness of advertising placements.

[0735] Input: Parsed emotion data

[0736] Output: Prediction of advertising effectiveness

[0737] The system predicts that advertising will be effective for topics that have a lot of positive sentiment, and prioritizes recommendations for those topics. If there are a lot of negative sentiments, the system decides not to advertise on those topics.

[0738] Step 8:

[0739] The server modifies and adjusts the ads based on the sentiment.

[0740] The server recommends changes to the content and creative of the advertisement based on the user's emotion recognition.

[0741] Input: Parsed emotion data

[0742] Output: Proposal for changes to ad content

[0743] For example, if you have a lot of positive emotions, we'll suggest ads with positive messages, and if you have a lot of negative emotions, we'll suggest ads with more calming messages.

[0744] (Application example 2)

[0745] 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."

[0746] Previous ad placement recommendation systems focused on predicting ad placements based on past data and current trend data. However, ad placements that do not take user emotions and reactions into account may not achieve the expected results. Furthermore, by taking real-time reactions into account, it is necessary to flexibly adjust the content and timing of ads. Therefore, the challenge is to use user emotion data to achieve more accurate ad placements and maximize advertising effectiveness.

[0747] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting past data, means for training a model based on the collected past data, means for collecting current trend data, means for predicting future popularity from the current trend data using the trained model, means for recommending advertisement placements based on the predicted popularity, means for collecting and analyzing user emotion data, and means for adjusting the content of advertisements based on the analyzed emotion data. This enables highly accurate advertisement placement that takes user emotions into consideration.

[0748] "Past data" refers to data about events or occurrences that occurred in the past, including viewership ratings, number of participants, number of posts on social media, number of related articles, etc.

[0749] "Current trend data" is data about events and content that are currently trending in society and the market, and is obtained from social media and news articles.

[0750] "Future popularity" is a value that predicts how much attention and interest a topic will attract in the future based on current trend data.

[0751] A "model" is a mathematical structure that is trained by a machine learning algorithm to learn patterns from given data and make predictions or classifications.

[0752] "Recommending advertising locations" means proposing the optimal location and timing for placing an ad based on predicted future popularity.

[0753] "User emotional data" refers to data that indicates the emotions and reactions expressed by users on social media or other platforms, including positive and negative emotions.

[0754] An "emotion engine" is software that collects and analyzes a user's emotional data to recognize their emotions.

[0755] "Features" are characteristics or attributes used by models to learn data in data analysis, including things like an event's viewership rating or number of attendees.

[0756] A "target variable" is a variable that is the target of prediction by a prediction model, and in this invention, future popularity is the target variable.

[0757] "Preprocessing" refers to a series of steps taken to convert raw data into an analyzable format, including data cleaning and transformation.

[0758] "Analysis" is the process of extracting useful information and patterns from collected data.

[0759] "Adjusting the content of the advertisement" means changing the message or creative of the advertisement based on the user's emotional data.

[0760] The present invention provides a method for implementing a system that optimizes advertising placement using user emotion data, and an embodiment thereof will be described in detail below.

[0761] The server collects current trend data from social media and news APIs. For example, it retrieves how often a particular hashtag is used or which topics are mentioned in news articles. This data is returned in JSON format and analyzed. Key keywords and temporal features are extracted during analysis. This process is performed using Python's requests library and data analysis libraries (e.g., Pandas).

[0762] Next, the server collects past event data, including viewership, number of participants, number of social media posts, number of related news articles, etc. The collected data is preprocessed to extract features and target variables. This is done using a machine learning library (e.g., scikit-learn).

[0763] The server trains a machine learning model based on the preprocessed historical data. The training process uses the features as input and outputs a popularity score. The trained model is then applied to a linear regression model or other regression models.

[0764] The server also uses current trend data to predict future popularity, allowing it to predict which events and content will attract attention in the future. The predicted popularity is output in JSON format, and a score is assigned to each trend.

[0765] Furthermore, the server collects and analyzes users' emotional data. A sentiment analysis engine is used to analyze users' social media posts and reactions. This engine typically uses natural language processing libraries (e.g., Hugging Face transformers).

[0766] The server applies an algorithm to adjust the content of ads based on the emotional data. For example, ads with positive messages are delivered to users based on positive emotional data, while ads with more neutral content are recommended for users based on negative emotional data. Existing ad management systems (e.g., Google Ads) can be used to adjust ads.

[0767] A concrete example of server processing is the 2023 Basketball World Cup. The server collects current topics from social media and news APIs and detects that the hashtag "BasketballWorldCup2023" is frequently used. It collects data on past sporting events and preprocesses their viewership and attendance figures. It then uses a machine learning model to predict future popularity and predicts that "BasketballWorldCup2023" will peak next weekend. It also collects sentiment data from users' social media posts, and if there are a lot of positive reactions, it delivers ads with positive messages, thereby increasing advertising effectiveness.

[0768] Examples of prompts include:

[0769] Collect current trend data and use it to predict future popularity. Additionally, collect user reactions from social media and analyze sentiment. Based on the results, recommend the best advertising destinations.

[0770] As described above, the system of the present invention provides a technique for utilizing user emotion data to realize effective and highly accurate advertising.

[0771] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0772] Step 1:

[0773] The server collects current trend data from social media and news APIs. Specifically, it sends requests to the APIs and obtains the returned JSON-formatted data. The input is the endpoints of the social media and news APIs, and the output is JSON-formatted data about current topics. A data analysis library (e.g., Pandas) is used to extract key keywords and temporal features.

[0774] Step 2:

[0775] The server collects past event data. It retrieves event viewership, number of participants, number of social media posts, number of related news articles, etc. from a database or another API. The input is queries or API requests related to past events, and the output is past event data. This data is preprocessed to extract features and target variables. Preprocessing is performed using a machine learning library (e.g., scikit-learn).

[0776] Step 3:

[0777] The server trains a machine learning model based on preprocessed historical data. The input is the preprocessed features and target variable, and the output is the trained model. The model can be a linear regression model or another regression model. A machine learning library (e.g., scikit-learn) is used.

[0778] Step 4:

[0779] The server uses current trend data to predict future popularity. The input is the current trend data and a trained model, and the output is the future popularity score for each trend. The preprocessed current trend data is input into the machine learning model to perform predictions. The prediction results are output in JSON format.

[0780] Step 5:

[0781] The server collects and analyzes user emotion data. The input is user posts and reaction data from social media, and the output is the analyzed emotion data. An emotion analysis engine (e.g., Hugging Face transformers) is used to recognize positive, negative, and neutral emotions in real time.

[0782] Step 6:

[0783] The server recommends ad placements based on future popularity scores and user emotion data. The input is future popularity scores and analyzed emotion data, and the output is a list of optimal ad placements. Connected to the ad management system, the recommendation engine suggests optimal ad placement locations and timing.

[0784] Step 7:

[0785] The server adjusts the content of the advertisements based on the emotional data. The input is the analyzed emotional data, and the output is the adjusted advertising content. If the user's emotional data is positive, advertisements with positive messages are delivered, and if the emotional data is negative, advertisements with gentle content are recommended. This process uses an existing advertising management system (e.g., Google Ads Management).

[0786] The above are the specific processing steps for carrying out the present invention.

[0787] 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.

[0788] 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.

[0789] 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.

[0790] [Third embodiment]

[0791] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0792] 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.

[0793] 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).

[0794] 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.

[0795] 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.

[0796] 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).

[0797] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for 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.

[0798] 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.

[0799] 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.

[0800] 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.

[0801] 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.

[0802] 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."

[0803] The present invention relates to a system for recommending advertisement placements, and includes processes for collecting past event data, training a model, collecting current trend data, predicting future popularity based on that data, and recommending advertisement placements. The program processing of this system is explained below in natural language.

[0804] Program Overview

[0805] The server retrieves the current trend

[0806] The server sends requests to social media and news APIs to retrieve current topics. This data, which contains information about specific topics or events, is returned in JSON format. The server analyzes this trend data to extract key keywords and temporal features.

[0807] The server collects and analyzes past event data

[0808] The server uses keywords extracted from current trends to collect past event data from a database or another API. This past data includes information about the popularity and attention of the event, such as viewer ratings, number of participants, number of social media posts, and number of related news articles. The server preprocesses this data and extracts features and target variables.

[0809] The server trains the model

[0810] The server uses the preprocessed historical event data to train a machine learning algorithm (e.g., a linear regression model). During this training process, features are input, popularity scores are output, and the model learns the relationship between them.

[0811] The server predicts future popularity

[0812] The server uses the trained model to predict future popularity from current trend data, allowing it to predict which events and content will attract attention in the future. The server adds the prediction results to the trend data and assigns a predicted popularity score to each trend.

[0813] The server recommends advertising

[0814] The server selects trends suitable for advertising based on the predicted popularity, prioritizes trends with high predicted popularity, and recommends the top trends that are expected to be particularly effective as advertising targets.

[0815] Specific examples

[0816] For example, if there is a sudden surge in excitement for the 2023 Basketball World Cup, the system would work as follows:

[0817] 1. The server collects current topics about the Basketball World Cup from social media and news APIs. For example, it detects that the hashtag "BasketballWorldCup2023" is frequently used.

[0818] 2. The server collects past sporting event data (e.g., the last Basketball World Cup, the Olympics) and preprocesses their viewership and attendance figures.

[0819] 3. The server uses this historical data to train a machine learning model (e.g., a linear regression model).

[0820] 4. The server uses the trained model to predict future popularity based on current trend data. The prediction reveals that the Basketball World Cup 2023 will peak next weekend.

[0821] 5. Based on the prediction results, the server recommends "Basketball World Cup 2023" as an advertising destination. It assigns a score indicating the applicability and priority of the advertising destination and recommends the optimal timing for advertising.

[0822] 6. After the ads are actually placed, the server collects the results and feeds them back into the model to improve prediction accuracy.

[0823] In this way, the system of the present invention can respond to trending content in real time and effectively recommend advertising, allowing advertisers to maximize their business opportunities without missing out.

[0824] The processing flow will be explained below.

[0825] Program processing steps

[0826] Step 1:

[0827] The server sends a request to a social networking or news API, such as the Twitter API or Google News API, to get the most popular topics at the moment, which allows you to get current trending data.

[0828] Step 2:

[0829] The server receives the API response and retrieves the trending data in JSON format, including specific keywords, topics, hashtags, etc.

[0830] Step 3:

[0831] The server analyzes the trend data and extracts keywords and temporal features (e.g., event dates and times, related hashtags) that are likely to be relevant to advertising. This data is used for subsequent model training.

[0832] Step 4:

[0833] The server then collects past event data from a database or other APIs based on the extracted keywords. For example, trending keywords related to the Basketball World Cup are used to collect data on similar past sporting events.

[0834] Step 5:

[0835] The server preprocesses past event data collected by the server. Specifically, it extracts features and target variables from data such as viewership, number of participants, and number of posts on social media. This creates a dataset in a format suitable for learning.

[0836] Step 6:

[0837] The server uses the preprocessed data to train a machine learning algorithm (e.g., a linear regression model). Through this training process, the model learns the relationship between features and popularity scores.

[0838] Step 7:

[0839] The server uses the trained model to predict future popularity based on current trend data. For example, it uses the current trend data of the Basketball World Cup as input and predicts future viewership and the likelihood of it becoming a hot topic on social media.

[0840] Step 8:

[0841] The server adds the predictions to the trend data and gives each trend a predicted popularity score, which reveals which trends are most suitable for advertising.

[0842] Step 9:

[0843] The server selects trends suitable for advertising based on predicted popularity, prioritizes trends with high predicted popularity, and recommends the top trends that are expected to be particularly effective as advertising targets.

[0844] Step 10:

[0845] The server notifies the user of the ad placement destinations. Specifically, it presents the user with a list of recommended ad placement destinations, along with details of the timing and location.

[0846] Step 11:

[0847] After the ad is placed, the server collects actual popularity data, such as viewer ratings and social media reactions after the ad is placed.

[0848] Step 12:

[0849] The server analyzes the difference between the actual popularity data and the predictions to identify areas for improvement in the model, providing feedback to improve the accuracy of the prediction model.

[0850] Step 13:

[0851] The server retrains the model using actual popularity data to improve prediction accuracy, enabling more accurate predictions for future ad placements.

[0852] Through the steps described above, the system of the present invention can respond to rapidly trending content in real time and effectively recommend advertising placements.

[0853] Example 1

[0854] 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."

[0855] Conventional advertising systems do not fully utilize past data and lack real-time trend analysis. As a result, they are unable to optimally select advertising destinations, resulting in reduced advertising effectiveness. Furthermore, they lack a feedback function to maintain and improve the accuracy of their predictive models, making it difficult to maximize advertising effectiveness over the long term.

[0856] 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.

[0857] In this invention, the server includes means for collecting past data, means for training a model based on the collected past data, means for collecting current trend data, means for predicting future popularity from the current trend data using the trained model, means for recommending advertising destinations based on the predicted popularity, means for acquiring and analyzing trend data from an API, means for preprocessing past data and extracting features and target variables, means for assigning popularity scores to each trend, and means for collecting actual results of advertising as feedback and improving the model. This enables advertisers to respond to rapidly trending content in real time and place effective advertisements.

[0858] "Past data" refers to information about past events, including viewership ratings, number of participants, number of posts on social media, and number of related news articles.

[0859] "Training a model" means using collected data to build a computational model using machine learning algorithms to learn patterns from the data and make predictions.

[0860] "Current trend data" is real-time information related to specific topics or events obtained from social media and news sites.

[0861] "Predicting future popularity" means using a trained model to predict the popularity of future events or topics based on current trend data.

[0862] "Recommending advertising destinations" means suggesting which events or topics you should advertise on based on predicted future popularity information.

[0863] "Obtaining and analyzing trend data from an API" means using an application programming interface (API) to obtain data from social media and news sites, and analyzing that data to extract useful information.

[0864] A "feature" is a specific attribute or characteristic of data that is used as input for a machine learning model.

[0865] A "target variable" is a value or category that is predicted as the output of a machine learning model.

[0866] A "Popularity Score" is a numerical representation of how much attention a particular event or topic will receive in the future.

[0867] "Feedback" refers to data collected from the results of actual advertising placements (e.g., click-through rate, conversion rate) and used to improve the performance of the model.

[0868] The present invention relates to a system for recommending advertisements, which trains a machine learning model based on past data and current trend data, predicts future popularity, and recommends advertisement placements. Specific embodiments for implementing this system are described below.

[0869] System configuration and technologies used

[0870] The server sends requests to the APIs of social media platforms and news services (e.g., Twitter API, Google News API) to obtain current trending data, which is returned in JSON format and parsed by the server to extract information about specific topics or events.

[0871] The server uses the extracted keywords to collect past data from a database or another API (e.g., sports database API, Google Trends API). The collected data includes viewership, number of participants, number of social media posts, number of related news articles, etc. The server preprocesses this data and extracts features and target variables. Specifically, it imputes missing values ​​and normalizes the data.

[0872] The server uses the preprocessed data to train a model using a machine learning algorithm (e.g., linear regression model or random forest). This training process uses the Scikit-learn library, with features as input and popularity scores as output. The dataset is split into training data and test data, and the performance of the model is evaluated.

[0873] The server uses the trained model to predict future popularity based on current trend data. Specifically, it inputs data from the "Basketball World Cup 2023" obtained from the Twitter API into the model and calculates the predicted level of attention for the following weekend. The output is a popularity score, which is used to evaluate future trends.

[0874] The server recommends optimal advertising destinations based on the predicted popularity score. For example, if interest in the "Basketball World Cup 2023" is predicted to be at its highest next weekend, the server will recommend this trend as an advertising destination. It also provides specific advertising strategies and timing to advertisers.

[0875] The server collects the results data (e.g., click-through rate, conversion rate) after the ads are actually placed and uses them as feedback to improve the accuracy of the model. The results data are used to retrain the model, maximizing the effectiveness of advertising in the long term.

[0876] Examples of specific examples and prompts

[0877] For example, if the 2023 Basketball World Cup is trending rapidly on social media, the system would work like this:

[0878] 1. The server collects current trend data related to "BasketballWorldCup2023" from the Twitter API and Google News API, and checks the frequency of the extracted keyword "BasketballWorldCup2023."

[0879] 2. The server collects viewership data and participant numbers from the previous Basketball World Cup from the sports database API, completes the missing data, and extracts it as features.

[0880] 3. The server uses Scikit-learn to train a linear regression model based on past data and build a model to predict the popularity of events.

[0881] 4. The server uses the trained model to predict future interest in the Basketball World Cup 2023 and estimates the peak interest for the following weekend.

[0882] 5. Based on the prediction results, the server recommends "Basketball World Cup 2023" as an advertising destination and proposes specific advertising timing and strategies to the advertiser.

[0883] 6. The server collects post-advertising result data (click-through rate and conversion rate) and retrains the predictive model to improve accuracy.

[0884] Example prompt sentence:

[0885] "Collect current social media trends related to the 2023 Basketball World Cup, use past sporting event data to predict future popularity, and recommend optimal advertising destinations. Specifically, retrieve trend data from the Twitter API, train a linear regression model using Scikit-learn, and propose an advertising strategy based on the prediction results."

[0886] In this way, the system of the present invention provides advertisers with a means to effectively place advertisements in response to rapidly trending content in real time.

[0887] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0888] Step 1:

[0889] The server collects current trend data. It sends requests to APIs of social media platforms and news services (e.g., Twitter API, Google News API) to retrieve data about current topics. The input is the API request, and the output is trend data in JSON format. The server parses this data to extract key information about specific topics or events, such as the frequency of occurrence of the hashtag "BasketballWorldCup2023."

[0890] Step 2:

[0891] The server collects past data. Using the keywords extracted in step 1, the server retrieves past related event data from a database or another API (e.g., sports database API, Google Trends API). This includes viewership, number of participants, number of social media posts, number of related news articles, etc. The input is keywords related to past events, and the output is a dataset of related past events. The server parses this dataset to extract the necessary information.

[0892] Step 3:

[0893] The server preprocesses the data. The server normalizes the past data collected in step 2, fills in missing values, and formats it as features and target variables. For example, it formats viewer ratings and number of posts as numerical data. The input is raw data of past events, and the output is a preprocessed dataset.

[0894] Step 4:

[0895] The server trains the model. The server uses the preprocessed data to train the model using a machine learning algorithm (e.g., linear regression model or random forest). For example, it uses the Scikit-learn library. The input is the preprocessed dataset, and the output is the trained predictive model. The model is configured to input features and output a popularity score.

[0896] Step 5:

[0897] The server predicts future popularity. The server uses the trained model to predict future popularity from the current trend data obtained in step 1. The input is the current trend data and the trained model, and the output is the future popularity score. For example, the server predicts that the hashtag "BasketballWorldCup2023" will peak next weekend.

[0898] Step 6:

[0899] The server recommends advertising destinations. The server selects the optimal advertising destination based on the popularity score predicted in step 5 and recommends advertising to the advertiser. The input is the future popularity score, and the output is the recommended advertising destination. Specifically, the server recommends "Basketball World Cup 2023" as an advertising destination and suggests advertising strategies and timing.

[0900] Step 7:

[0901] The server collects feedback and improves the model. The server collects results after the ads are actually placed (e.g., click-through rate, conversion rate), and uses that data to retrain the model and improve its accuracy. The input is the results data after the ads are placed, and the output is an improved model. This makes it possible to maximize advertising effectiveness in the long term.

[0902] (Application example 1)

[0903] 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."

[0904] Conventional advertising systems statically select targets based on past data and are unable to respond to real-time trend changes, making it difficult to optimize advertising effectiveness. Furthermore, there is no feedback based on post-advertising results to improve model accuracy, meaning long-term improvements in advertising effectiveness cannot be expected. Furthermore, there is a lack of a way for users to easily evaluate advertising effectiveness and reflect that in their next ad placement.

[0905] 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.

[0906] In this invention, the server includes means for collecting past event data, means for training a model based on the collected past event data, means for collecting current trend data, means for predicting future popularity from the current trend data using the trained model, means for recommending advertising destinations based on the predicted popularity, means for users to provide feedback on the results of their advertising placements, and means for improving the accuracy of the model using the feedback results. This enables rapid response to trend fluctuations in real time and optimization of advertising effectiveness. Furthermore, improving the accuracy of the model based on user feedback can also be expected to improve long-term advertising effectiveness. Furthermore, users can easily evaluate advertising effectiveness and reflect it in their next advertising placements.

[0907] "Past event data" refers to information related to events that have taken place in the past, including viewership ratings, number of participants, number of posts on social media, number of related news articles, and the like.

[0908] "Training a model" is the process of using machine learning algorithms to learn relationships between collected data.

[0909] "Current trending data" is data on information and topics that are currently trending, collected from social media, news, and other sources.

[0910] "Predicting future popularity" is the process of using a trained model to infer what will be popular and trending in the future based on current trend data.

[0911] "Recommending advertising locations" means suggesting the most effective locations and timing for advertising based on predicted popularity.

[0912] "Providing feedback" means that the user reports to the server the results and effects obtained after placing an advertisement.

[0913] "Improving the accuracy of the model" means adjusting the parameters and algorithms of the predictive model based on the feedback results to increase the accuracy of the next prediction.

[0914] This invention relates to a system for recommending advertising, which consists of the following steps: First, a server collects past event data and trains a machine learning model based on this data. Next, the server collects current trend data and uses the trained model to predict future popularity. Based on the prediction results, the system recommends optimal advertising destinations to users.

[0915] Program processing

[0916] Data Collection Module

[0917] The server collects current trend data via social media and news APIs. Specifically, it uses Python to send requests to the Twitter API and Google News API, and retrieves JSON-formatted data including trending keywords and the number of posts.

[0918] Hardware used: Server, smartphone

[0919] Software used: Python, Twitter API, Google News API

[0920] Historical Data Analysis Module

[0921] The server collects past event data from a database or another API, and performs preprocessing to extract features and target variables. In this process, metrics such as viewership, number of participants, and number of posts on social media are used, and the data is organized using Python's Scikit-learn.

[0922] Hardware used: Server

[0923] Software used: Python, Scikit-learn

[0924] Model Training Module

[0925] The server uses the preprocessed data to train a machine learning algorithm (e.g., a linear regression model), which learns relationships in the data and generates a model to predict future popularity.

[0926] Hardware used: Server

[0927] Software used: Python, Scikit-learn

[0928] Popularity Prediction Module

[0929] The server uses the trained model to predict future popularity based on current trend data, a process also performed using Python.

[0930] Hardware used: Server

[0931] Software used: Python, Scikit-learn

[0932] Recommendation Engine

[0933] The server recommends the most effective advertising destinations based on predicted popularity, allowing users to know effective advertising destinations in real time.

[0934] Hardware used: Server

[0935] Software used: Python, Flask

[0936] Feedback Module

[0937] Users provide feedback on the results of their ad placements, and the data is sent to the server, where it is used to retrain the model, improving its prediction accuracy.

[0938] Hardware used: Smartphone

[0939] Software used: React Native, Firebase

[0940] Specific examples

[0941] As a specific application example, we will explain the process in which a server collects the latest trend data and predicts future popularity. For example, the server inputs the following prompt sentence into the generative AI model:

[0942] Gather the latest trends and predict future popularity. Use past event data to identify hot topics for the upcoming weekend and suggest them for advertising.

[0943] Through this process, the server can quickly respond to trend fluctuations in real time and optimize advertising effectiveness. It also has a feedback function to improve the accuracy of the model, making it a system that can achieve long-term improvements in advertising effectiveness.

[0944] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0945] Step 1:

[0946] The server collects past event data. Specifically, the server retrieves data related to past events, such as viewer ratings, number of participants, and number of posts on social media, from a database or API. After collection, this data is saved in JSON format and used for subsequent processing.

[0947] Input: Historical event data from a database or API

[0948] Output: Historical event data stored in JSON format

[0949] Step 2:

[0950] The server preprocesses the collected past event data and extracts features and target variables. Specifically, it performs missing value imputation, noise removal, and scaling of the data, and then extracts features (e.g., viewership rate, number of participants, number of social media posts) and target variables. Python's Scikit-learn is used.

[0951] Input: Historical event data stored in JSON format

[0952] Output: Preprocessed features and target variable

[0953] Step 3:

[0954] The server trains the model using the preprocessed data. Specifically, it uses a machine learning algorithm such as a linear regression model, inputs features, and outputs target variables. Python's Scikit-learn is used.

[0955] Input: Preprocessed features and target variables

[0956] Output: A trained machine learning model

[0957] Step 4:

[0958] The server collects current trend data from social media and news APIs. Specifically, it sends an API request and receives data in JSON format containing the latest topics and keywords. For example, it uses the Twitter API or Google News API.

[0959] Input: Trend data from Twitter API and Google News API

[0960] Output: Current trend data saved in JSON format

[0961] Step 5:

[0962] The server preprocesses the collected current trend data and extracts features. Specifically, it performs noise removal and normalization, and extracts features (e.g., topics, keywords, and number of posts). It uses Python's Scikit-learn.

[0963] Input: Current trend data stored in JSON format

[0964] Output: Preprocessed features

[0965] Step 6:

[0966] The server uses the trained model to predict future popularity based on current trend data. Specifically, the preprocessed features are input into the model to predict future popularity scores.

[0967] Input: Preprocessed features, trained model

[0968] Output: Predicted future popularity scores

[0969] Step 7:

[0970] The server recommends advertising destinations based on the predicted popularity. Specifically, it prioritizes trends with high popularity scores and suggests them as advertising destinations. Users receive this information through a smartphone application.

[0971] Input: Predicted future popularity scores

[0972] Output: Recommended ad placements

[0973] Step 8:

[0974] Users provide feedback on the results of their ads after placing them. Specifically, they use their smartphones to input advertising effectiveness measurement data (click rates, conversion rates, etc.) and send it to the server.

[0975] Input: Post-advertising effectiveness measurement data

[0976] Output: Feedback data sent to the server

[0977] Step 9:

[0978] The server uses the feedback data to improve the accuracy of the model. Specifically, it uses the feedback data as retraining data to adjust the model parameters and algorithms, thereby improving the accuracy of the next prediction.

[0979] Input: Feedback data

[0980] Output: A trained model with improved accuracy

[0981] 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.

[0982] The present invention combines an advertising recommendation system with an emotion engine that recognizes user emotions, and in addition to the process of collecting past event data, training a model based on this, collecting current trend data to predict future popularity, and recommending advertising destinations, the system utilizes user emotion data to achieve more accurate advertising placement. The program processing of this system is explained below in natural language.

[0983] Program Overview

[0984] The server retrieves the current trend

[0985] The server sends requests to social media and news APIs to retrieve current topics. This data, which contains information about specific topics or events, is returned in JSON format. The server analyzes this trend data to extract key keywords and temporal features.

[0986] The server collects and analyzes past event data

[0987] The server uses keywords extracted from current trends to collect past event data from a database or another API. This past data includes information about the popularity and attention of the event, such as viewer ratings, number of participants, number of social media posts, and number of related news articles. The server preprocesses this data and extracts features and target variables.

[0988] The server trains the model

[0989] The server uses the preprocessed historical event data to train a machine learning algorithm (e.g., a linear regression model). During this training process, features are input, popularity scores are output, and the model learns the relationship between them.

[0990] The server predicts future popularity

[0991] The server uses the trained model to predict future popularity from current trend data, allowing it to predict which events and content will attract attention in the future. The server adds the prediction results to the trend data and assigns a predicted popularity score to each trend.

[0992] The server recommends advertising

[0993] The server selects trends suitable for advertising based on the predicted popularity, prioritizes trends with high predicted popularity, and recommends the top trends that are expected to be particularly effective as advertising targets.

[0994] Introducing the Emotion Engine

[0995] The server uses an emotion engine to recognize the user's emotions. Specifically, it collects and analyzes the user's emotional data (e.g., comments and reactions posted on social media) to understand the user's emotional state in real time.

[0996] Analyzing Emotional Data

[0997] The server analyzes the collected emotion data and predicts the effectiveness of advertising. For example, it predicts that advertising will be more effective for topics that receive a lot of positive emotion, and prioritizes recommendations for those topics. Conversely, if there are a lot of negative emotions, it decides not to advertise on those topics.

[0998] Advertisement changes and adjustments based on emotions

[0999] The server can also recommend changes to ad content or creative based on user emotion recognition. For example, if a user's emotions are positive, it can suggest ads with positive messages. Conversely, if a user's emotions are predominantly negative, it can suggest ads with more neutral content.

[1000] Specific examples

[1001] For example, the following actions are taken for the 2023 Basketball World Cup:

[1002] 1. The server collects current topics from social media and news APIs:

[1003] The server sends a request to the API to retrieve trending data about the Basketball World Cup. For example, it detects that the hashtag "BasketballWorldCup2023" is popular.

[1004] 2. The server collects and analyzes past event data:

[1005] The server collects data on past sporting events (e.g., the last Basketball World Cup, the Olympics) and preprocesses their viewership and attendance figures.

[1006] 3. The server trains the model:

[1007] The server uses this historical data to train a machine learning model (e.g., a linear regression model).

[1008] 4. The server predicts future popularity:

[1009] The server uses the trained model to predict future popularity based on current trend data, and the prediction reveals that the Basketball World Cup 2023 will peak next weekend.

[1010] 5. The server recommends advertising:

[1011] Based on the prediction results, the server recommends "Basketball World Cup 2023" as an advertising destination. It assigns a score indicating the applicability and priority of the advertising destination and recommends the optimal timing for advertising.

[1012] 6. Introducing the Emotion Engine:

[1013] The server collects user reactions and comments from social media and analyzes them to understand user sentiment. For example, if the reaction to the "Basketball World Cup 2023" is positive, it determines that advertising will be effective.

[1014] 7. Advertisement modification and adjustment based on emotions:

[1015] If the server detects a user's positive emotions, it will suggest ads with positive messages. Conversely, if a user's negative emotions are detected, it will suggest more neutral ads.

[1016] In this way, the system of the present invention can not only respond to rapidly trending content in real time and effectively recommend advertising placements, but also take user emotions into consideration, thereby achieving even more effective advertising operations.

[1017] The processing flow will be explained below.

[1018] Step 1:

[1019] The server sends a request to a social media or news API to retrieve the most popular topics. Specifically, it makes a request to the Twitter API or Google News API to retrieve current trending data. The returned data is received in JSON format.

[1020] Step 2:

[1021] The server parses the API response and converts the trend data from JSON format into Python objects, making it easier to work with in your programs.

[1022] Step 3:

[1023] The server extracts keywords and temporal features (e.g., event dates and times, related hashtags) that are likely to be relevant to advertising from the trend data it acquires. This extraction is performed through text analysis of the trend data.

[1024] Step 4:

[1025] Based on the keywords extracted by the server, past event data is collected from a database or another API, and data on past sporting and entertainment events related to the extracted keywords is obtained.

[1026] Step 5:

[1027] The server preprocesses the collected historical event data. This preprocessing involves extracting features and target variables from data such as viewership, number of participants, and number of posts on social media. This prepares the data in a format suitable for model training.

[1028] Step 6:

[1029] The server uses the preprocessed data to train a machine learning algorithm (e.g., a linear regression model). Through the training process, the model learns the relationship between features and popularity scores.

[1030] Step 7:

[1031] The server uses the trained model to predict future popularity based on current trend data. Specifically, current trend data is input into the model, and predicted viewership ratings and social media reactions are output.

[1032] Step 8:

[1033] The server adds the predicted popularity score to the trend data, which gives each trend a numerical popularity score that can be used to determine advertising placements.

[1034] Step 9:

[1035] The server selects the best trend for advertising based on the popularity score. Specifically, it prioritizes trends with high popularity scores and lists them as recommended advertising destinations.

[1036] Step 10:

[1037] The server obtains user reactions and comments from the social media platform to collect user sentiment data, which includes the process of collecting reaction data through APIs.

[1038] Step 11:

[1039] The server analyzes the collected emotional data to identify the user's emotional state. Sentiment analysis uses natural language processing technology to classify emotions such as positive and negative from the content of comments.

[1040] Step 12:

[1041] The server changes or adjusts advertising destinations based on the emotion data. For example, it determines that topics with a high percentage of positive emotions have a high advertising effectiveness and recommends them as priority destinations.

[1042] Step 13:

[1043] The server combines emotion data with trend data to improve the accuracy of the model, enabling more accurate predictions that take into account the impact of emotional states on predictions.

[1044] Step 14:

[1045] The server notifies the user of recommended advertising placements. Specifically, it presents the user with a list of recommended advertising placements and provides detailed information on the optimal timing and location.

[1046] Step 15:

[1047] After the ad is placed, the server collects actual ad performance data, including viewer ratings, social media responses, and click counts.

[1048] Step 16:

[1049] The server analyzes the performance data collected and analyzes the difference between predicted and actual results, which allows the model to identify areas for improvement.

[1050] Step 17:

[1051] The server uses the feedback to retrain the model and improve its prediction accuracy. By incorporating real data, future predictions will be more accurate.

[1052] Through the above steps, the system of the present invention can respond to rapidly trending content in real time and realize effective advertising that takes into account user emotions.

[1053] Example 2

[1054] 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."

[1055] Conventional advertising systems typically place ads based on past data and current trend data. However, these systems do not take into account user emotional data, which limits the effectiveness of advertising and makes it difficult to optimally place ads. Particularly in today's world, where real-time data such as social media is increasingly being used, ignoring user emotions and reactions will result in a decrease in advertising effectiveness. Therefore, there is a demand for a system that takes user emotions into account and enables more accurate advertising placement.

[1056] 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 a means for collecting past event data, a means for training a model based on the collected past event data, and a means for collecting current trend data. This enables a means for collecting and analyzing user emotion data and a means for adjusting the content and creative of advertisements based on the emotion data.

[1057] "Past event data" refers to information such as viewership ratings, number of participants, number of social media posts, and number of related news articles related to a specific period or event.

[1058] "Training a model" refers to using collected data to train a machine learning algorithm to learn the relationship between features and target variables, making it usable for future predictions.

[1059] "Current trend data" refers to information about the latest topics and trends collected from social media and news APIs.

[1060] "Predicting future popularity" refers to using a trained model to calculate future attention and popularity from current trend data.

[1061] "Recommending advertising destinations" refers to suggesting optimal advertising investment destinations based on predicted popularity.

[1062] "User emotional data" refers to information on emotions and reactions obtained by analyzing comments and reactions posted by users on social media, etc.

[1063] "Collecting and analyzing emotional data" refers to collecting user comments and reactions and analyzing them to understand their emotional state, whether positive, negative, or neutral.

[1064] "Adjusting advertising content and creative based on emotional data" refers to optimizing advertising messages and visuals based on collected and analyzed emotional data.

[1065] This invention combines an advertising recommendation system with an emotion engine that recognizes user emotions. Specifically, it collects past event data, trains a model based on this, collects current trend data to predict future popularity, and recommends advertising destinations. In addition, by utilizing user emotion data, it achieves more accurate advertising placement.

[1066] The server sends requests to social media and news APIs (e.g., Twitter API, NewsAPI) to gather current topics. The data is often returned in JSON format, and the server parses it to extract keywords related to a particular topic or event. For example, the keyword "2023 Basketball World Cup" might be extracted.

[1067] Next, the server uses the keywords extracted from the current trends to collect past event data from a database or another API. This data includes detailed information about the popularity and attention of the event, such as viewership, number of participants, number of social media posts, and number of related news articles. The server preprocesses this data and extracts features and target variables.

[1068] The server then uses the preprocessed historical event data to train a machine learning algorithm (e.g., a linear regression model). Through this process, the model learns the relationship between features and popularity scores. Once trained, the model is used to predict future popularity from current trend data. This prediction process makes it possible to predict which events or content will attract attention in the future. The server adds the prediction results to the trend data and assigns a predicted popularity score to each trend.

[1069] Next, the server selects trends suitable for advertising based on the predicted popularity. For example, it prioritizes popular trends and recommends the top trends that are expected to be particularly effective as advertising targets.

[1070] Furthermore, the server will incorporate an emotion engine to collect and analyze user emotional data. Specifically, it will collect user reactions and comments from social media and analyze them to understand the user's emotional state in real time. For example, it will use IBM Watson's Natural Language Understanding API. This analysis will enable it to predict the effectiveness of advertising, predicting that advertising will be more effective for topics that attract a lot of positive emotions and prioritizing those topics for recommendations. Conversely, if there are a lot of negative emotions, it will decide not to advertise.

[1071] Furthermore, the server can recommend changes to the content and creative of advertisements based on the user's emotional state. For example, if the user's emotional state is positive, the server can suggest advertisements with positive messages. Conversely, if the user's emotional state is mostly negative, the server can suggest advertisements with more neutral content.

[1072] As a concrete example, for the 2023 Basketball World Cup, the following actions are taken:

[1073] The server uses the Twitter API to obtain trend data related to "BasketballWorldCup2023." For example, it detects that the hashtag "BasketballWorldCup2023" is frequently used. Next, the server collects and preprocesses viewership and attendance data related to previous Basketball World Cups and Olympics from Google Analytics. Next, the server trains a linear regression model using this historical data. Using the trained model, it predicts future popularity based on current trend data and predicts that "BasketballWorldCup2023" will peak next weekend. Based on the prediction results, the server then recommends "BasketballWorldCup2023" as an advertising destination. Next, the server uses IBM Watson's API to confirm positive sentiment toward "BasketballWorldCup2023" and suggests ads containing positive messages.

[1074] In this way, the system of the present invention can not only respond to rapidly trending content in real time and effectively recommend advertising placements, but also take user emotions into consideration, thereby achieving even more effective advertising operations.

[1075] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1076] Step 1:

[1077] The server retrieves the current trend.

[1078] The server sends requests to social media and news APIs (e.g., Twitter API, News API) to collect current topics. For example, to obtain trending data about the "Basketball World Cup 2023," the server makes the following API request:

[1079] Input: API request for "BasketballWorldCup2023"

[1080] Output: Trend data in JSON format

[1081] The server parses this JSON data and extracts keywords related to specific topics or events, using Python libraries to parse the JSON data and extract the necessary information.

[1082] Step 2:

[1083] The server collects and analyzes past event data.

[1084] The server uses keywords extracted from current trends and gathers data on past events from a database or another API, for example, data on past Basketball World Cups or Olympic Games.

[1085] Input: Current trending keyword (e.g. "Basketball World Cup")

[1086] Output: Past event data (e.g., viewership, number of participants, number of social media posts)

[1087] The server preprocesses this data and extracts features (e.g., viewership rate, number of SNS posts) and target variables (e.g., popularity score) using Python's pandas library.

[1088] Step 3:

[1089] The server trains the model.

[1090] The server uses the pre-processed historical event data to train a machine learning algorithm (e.g., a linear regression model).

[1091] Input: Preprocessed historical event data (features and target variables)

[1092] Output: The trained model

[1093] Implement and train a linear regression model using Python's scikit-learn library.

[1094] Step 4:

[1095] The server predicts future popularity.

[1096] The server uses the trained model to predict future popularity from current trend data.

[1097] Input: Features of the current trend data

[1098] Output: Predicted popularity score

[1099] The trained model is used to make predictions and assign each trend a predicted popularity score.

[1100] Step 5:

[1101] The server recommends advertising placements.

[1102] The server selects trends suitable for advertising based on the predicted popularity.

[1103] Input: Predicted popularity score

[1104] Output: A list of recommended ad placements

[1105] Based on the prediction results, the most popular trends are prioritized, and the top trends that are expected to be particularly effective are recommended as advertising targets.

[1106] Step 6:

[1107] The server implements the emotion engine.

[1108] The server uses an emotion engine to collect and analyze user emotion data, for example, by collecting user reactions and comments from social media and analyzing their emotions.

[1109] Input: Social media comments and reactions

[1110] Output: Parsed sentiment data (positive, negative, neutral, etc.)

[1111] Sentiment analysis is performed using IBM Watson's Natural Language Understanding API, among others.

[1112] Step 7:

[1113] The server analyzes the emotion data.

[1114] The server analyzes the collected emotional data and predicts the effectiveness of advertising placements.

[1115] Input: Parsed emotion data

[1116] Output: Prediction of advertising effectiveness

[1117] The system predicts that advertising will be effective for topics that have a lot of positive sentiment, and prioritizes recommendations for those topics. If there are a lot of negative sentiments, the system decides not to advertise on those topics.

[1118] Step 8:

[1119] The server modifies and adjusts the ads based on the sentiment.

[1120] The server recommends changes to the content and creative of the advertisement based on the user's emotion recognition.

[1121] Input: Parsed emotion data

[1122] Output: Proposal for changes to ad content

[1123] For example, if you have a lot of positive emotions, we'll suggest ads with positive messages, and if you have a lot of negative emotions, we'll suggest ads with more calming messages.

[1124] (Application example 2)

[1125] 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."

[1126] Previous ad placement recommendation systems focused on predicting ad placements based on past data and current trend data. However, ad placements that do not take user emotions and reactions into account may not achieve the expected results. Furthermore, by taking real-time reactions into account, it is necessary to flexibly adjust the content and timing of ads. Therefore, the challenge is to use user emotion data to achieve more accurate ad placements and maximize advertising effectiveness.

[1127] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting past data, means for training a model based on the collected past data, means for collecting current trend data, means for predicting future popularity from the current trend data using the trained model, means for recommending advertisement placements based on the predicted popularity, means for collecting and analyzing user emotion data, and means for adjusting the content of advertisements based on the analyzed emotion data. This enables highly accurate advertisement placement that takes user emotions into consideration.

[1128] "Past data" refers to data about events or occurrences that occurred in the past, including viewership ratings, number of participants, number of posts on social media, number of related articles, etc.

[1129] "Current trend data" is data about events and content that are currently trending in society and the market, and is obtained from social media and news articles.

[1130] "Future popularity" is a value that predicts how much attention and interest a topic will attract in the future based on current trend data.

[1131] A "model" is a mathematical structure that is trained by a machine learning algorithm to learn patterns from given data and make predictions or classifications.

[1132] "Recommending advertising locations" means proposing the optimal location and timing for placing an ad based on predicted future popularity.

[1133] "User emotional data" refers to data that indicates the emotions and reactions expressed by users on social media or other platforms, including positive and negative emotions.

[1134] An "emotion engine" is software that collects and analyzes a user's emotional data to recognize their emotions.

[1135] "Features" are characteristics or attributes used by models to learn data in data analysis, including things like an event's viewership rating or number of attendees.

[1136] A "target variable" is a variable that is the target of prediction by a prediction model, and in this invention, future popularity is the target variable.

[1137] "Preprocessing" refers to a series of steps taken to convert raw data into an analyzable format, including data cleaning and transformation.

[1138] "Analysis" is the process of extracting useful information and patterns from collected data.

[1139] "Adjusting the content of the advertisement" means changing the message or creative of the advertisement based on the user's emotional data.

[1140] The present invention provides a method for implementing a system that optimizes advertising placement using user emotion data, and an embodiment thereof will be described in detail below.

[1141] The server collects current trend data from social media and news APIs. For example, it retrieves how often a particular hashtag is used or which topics are mentioned in news articles. This data is returned in JSON format and analyzed. Key keywords and temporal features are extracted during analysis. This process is performed using Python's requests library and data analysis libraries (e.g., Pandas).

[1142] Next, the server collects past event data, including viewership, number of participants, number of social media posts, number of related news articles, etc. The collected data is preprocessed to extract features and target variables. This is done using a machine learning library (e.g., scikit-learn).

[1143] The server trains a machine learning model based on the preprocessed historical data. The training process uses the features as input and outputs a popularity score. The trained model is then applied to a linear regression model or other regression models.

[1144] The server also uses current trend data to predict future popularity, allowing it to predict which events and content will attract attention in the future. The predicted popularity is output in JSON format, and a score is assigned to each trend.

[1145] Furthermore, the server collects and analyzes users' emotional data. A sentiment analysis engine is used to analyze users' social media posts and reactions. This engine typically uses natural language processing libraries (e.g., Hugging Face transformers).

[1146] The server applies an algorithm to adjust the content of ads based on the emotional data. For example, ads with positive messages are delivered to users based on positive emotional data, while ads with more neutral content are recommended for users based on negative emotional data. Existing ad management systems (e.g., Google Ads) can be used to adjust ads.

[1147] A concrete example of server processing is the 2023 Basketball World Cup. The server collects current topics from social media and news APIs and detects that the hashtag "BasketballWorldCup2023" is frequently used. It collects data on past sporting events and preprocesses their viewership and attendance figures. It then uses a machine learning model to predict future popularity and predicts that "BasketballWorldCup2023" will peak next weekend. It also collects sentiment data from users' social media posts, and if there are a lot of positive reactions, it delivers ads with positive messages, thereby increasing advertising effectiveness.

[1148] Examples of prompts include:

[1149] Collect current trend data and use it to predict future popularity. Additionally, collect user reactions from social media and analyze sentiment. Based on the results, recommend the best advertising destinations.

[1150] As described above, the system of the present invention provides a technique for utilizing user emotion data to realize effective and highly accurate advertising.

[1151] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1152] Step 1:

[1153] The server collects current trend data from social media and news APIs. Specifically, it sends requests to the APIs and obtains the returned JSON-formatted data. The input is the endpoints of the social media and news APIs, and the output is JSON-formatted data about current topics. A data analysis library (e.g., Pandas) is used to extract key keywords and temporal features.

[1154] Step 2:

[1155] The server collects past event data. It retrieves event viewership, number of participants, number of social media posts, number of related news articles, etc. from a database or another API. The input is queries or API requests related to past events, and the output is past event data. This data is preprocessed to extract features and target variables. Preprocessing is performed using a machine learning library (e.g., scikit-learn).

[1156] Step 3:

[1157] The server trains a machine learning model based on preprocessed historical data. The input is the preprocessed features and target variable, and the output is the trained model. The model can be a linear regression model or another regression model. A machine learning library (e.g., scikit-learn) is used.

[1158] Step 4:

[1159] The server uses current trend data to predict future popularity. The input is the current trend data and a trained model, and the output is the future popularity score for each trend. The preprocessed current trend data is input into the machine learning model to perform predictions. The prediction results are output in JSON format.

[1160] Step 5:

[1161] The server collects and analyzes user emotion data. The input is user posts and reaction data from social media, and the output is the analyzed emotion data. An emotion analysis engine (e.g., Hugging Face transformers) is used to recognize positive, negative, and neutral emotions in real time.

[1162] Step 6:

[1163] The server recommends ad placements based on future popularity scores and user emotion data. The input is future popularity scores and analyzed emotion data, and the output is a list of optimal ad placements. Connected to the ad management system, the recommendation engine suggests optimal ad placement locations and timing.

[1164] Step 7:

[1165] The server adjusts the content of the advertisements based on the emotional data. The input is the analyzed emotional data, and the output is the adjusted advertising content. If the user's emotional data is positive, advertisements with positive messages are delivered, and if the emotional data is negative, advertisements with gentle content are recommended. This process uses an existing advertising management system (e.g., Google Ads Management).

[1166] The above are the specific processing steps for carrying out the present invention.

[1167] 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.

[1168] 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.

[1169] 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.

[1170] [Fourth embodiment]

[1171] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1172] 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.

[1173] 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).

[1174] 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.

[1175] 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.

[1176] 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).

[1177] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for 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.

[1178] 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.

[1179] 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.

[1180] 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.

[1181] 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.

[1182] 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.

[1183] 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."

[1184] The present invention relates to a system for recommending advertisement placements, and includes processes for collecting past event data, training a model, collecting current trend data, predicting future popularity based on that data, and recommending advertisement placements. The program processing of this system is explained below in natural language.

[1185] Program Overview

[1186] The server retrieves the current trend

[1187] The server sends requests to social media and news APIs to retrieve current topics. This data, which contains information about specific topics or events, is returned in JSON format. The server analyzes this trend data to extract key keywords and temporal features.

[1188] The server collects and analyzes past event data

[1189] The server uses keywords extracted from current trends to collect past event data from a database or another API. This past data includes information about the popularity and attention of the event, such as viewer ratings, number of participants, number of social media posts, and number of related news articles. The server preprocesses this data and extracts features and target variables.

[1190] The server trains the model

[1191] The server uses the preprocessed historical event data to train a machine learning algorithm (e.g., a linear regression model). During this training process, features are input, popularity scores are output, and the model learns the relationship between them.

[1192] The server predicts future popularity

[1193] The server uses the trained model to predict future popularity from current trend data, allowing it to predict which events and content will attract attention in the future. The server adds the prediction results to the trend data and assigns a predicted popularity score to each trend.

[1194] The server recommends advertising

[1195] The server selects trends suitable for advertising based on the predicted popularity, prioritizes trends with high predicted popularity, and recommends the top trends that are expected to be particularly effective as advertising targets.

[1196] Specific examples

[1197] For example, if there is a sudden surge in excitement for the 2023 Basketball World Cup, the system would work as follows:

[1198] 1. The server collects current topics about the Basketball World Cup from social media and news APIs. For example, it detects that the hashtag "BasketballWorldCup2023" is frequently used.

[1199] 2. The server collects past sporting event data (e.g., the last Basketball World Cup, the Olympics) and preprocesses their viewership and attendance figures.

[1200] 3. The server uses this historical data to train a machine learning model (e.g., a linear regression model).

[1201] 4. The server uses the trained model to predict future popularity based on current trend data. The prediction reveals that the Basketball World Cup 2023 will peak next weekend.

[1202] 5. Based on the prediction results, the server recommends "Basketball World Cup 2023" as an advertising destination. It assigns a score indicating the applicability and priority of the advertising destination and recommends the optimal timing for advertising.

[1203] 6. After the ads are actually placed, the server collects the results and feeds them back into the model to improve prediction accuracy.

[1204] In this way, the system of the present invention can respond to trending content in real time and effectively recommend advertising, allowing advertisers to maximize their business opportunities without missing out.

[1205] The processing flow will be explained below.

[1206] Program processing steps

[1207] Step 1:

[1208] The server sends a request to a social networking or news API, such as the Twitter API or Google News API, to get the most popular topics at the moment, which allows you to get current trending data.

[1209] Step 2:

[1210] The server receives the API response and retrieves the trending data in JSON format, including specific keywords, topics, hashtags, etc.

[1211] Step 3:

[1212] The server analyzes the trend data and extracts keywords and temporal features (e.g., event dates and times, related hashtags) that are likely to be relevant to advertising. This data is used for subsequent model training.

[1213] Step 4:

[1214] The server then collects past event data from a database or other APIs based on the extracted keywords. For example, trending keywords related to the Basketball World Cup are used to collect data on similar past sporting events.

[1215] Step 5:

[1216] The server preprocesses past event data collected by the server. Specifically, it extracts features and target variables from data such as viewership, number of participants, and number of posts on social media. This creates a dataset in a format suitable for learning.

[1217] Step 6:

[1218] The server uses the preprocessed data to train a machine learning algorithm (e.g., a linear regression model). Through this training process, the model learns the relationship between features and popularity scores.

[1219] Step 7:

[1220] The server uses the trained model to predict future popularity based on current trend data. For example, it uses the current trend data of the Basketball World Cup as input and predicts future viewership and the likelihood of it becoming a hot topic on social media.

[1221] Step 8:

[1222] The server adds the predictions to the trend data and gives each trend a predicted popularity score, which reveals which trends are most suitable for advertising.

[1223] Step 9:

[1224] The server selects trends suitable for advertising based on predicted popularity, prioritizes trends with high predicted popularity, and recommends the top trends that are expected to be particularly effective as advertising targets.

[1225] Step 10:

[1226] The server notifies the user of the ad placement destinations. Specifically, it presents the user with a list of recommended ad placement destinations, along with details of the timing and location.

[1227] Step 11:

[1228] After the ad is placed, the server collects actual popularity data, such as viewer ratings and social media reactions after the ad is placed.

[1229] Step 12:

[1230] The server analyzes the difference between the actual popularity data and the predictions to identify areas for improvement in the model, providing feedback to improve the accuracy of the prediction model.

[1231] Step 13:

[1232] The server retrains the model using actual popularity data to improve prediction accuracy, enabling more accurate predictions for future ad placements.

[1233] Through the steps described above, the system of the present invention can respond to rapidly trending content in real time and effectively recommend advertising placements.

[1234] Example 1

[1235] 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."

[1236] Conventional advertising systems do not fully utilize past data and lack real-time trend analysis. As a result, they are unable to optimally select advertising destinations, resulting in reduced advertising effectiveness. Furthermore, they lack a feedback function to maintain and improve the accuracy of their predictive models, making it difficult to maximize advertising effectiveness over the long term.

[1237] 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.

[1238] In this invention, the server includes means for collecting past data, means for training a model based on the collected past data, means for collecting current trend data, means for predicting future popularity from the current trend data using the trained model, means for recommending advertising destinations based on the predicted popularity, means for acquiring and analyzing trend data from an API, means for preprocessing past data and extracting features and target variables, means for assigning popularity scores to each trend, and means for collecting actual results of advertising as feedback and improving the model. This enables advertisers to respond to rapidly trending content in real time and place effective advertisements.

[1239] "Past data" refers to information about past events, including viewership ratings, number of participants, number of posts on social media, and number of related news articles.

[1240] "Training a model" means using collected data to build a computational model using machine learning algorithms to learn patterns from the data and make predictions.

[1241] "Current trend data" is real-time information related to specific topics or events obtained from social media and news sites.

[1242] "Predicting future popularity" means using a trained model to predict the popularity of future events or topics based on current trend data.

[1243] "Recommending advertising destinations" means suggesting which events or topics you should advertise on based on predicted future popularity information.

[1244] "Obtaining and analyzing trend data from an API" means using an application programming interface (API) to obtain data from social media and news sites, and analyzing that data to extract useful information.

[1245] A "feature" is a specific attribute or characteristic of data that is used as input for a machine learning model.

[1246] A "target variable" is a value or category that is predicted as the output of a machine learning model.

[1247] A "Popularity Score" is a numerical representation of how much attention a particular event or topic will receive in the future.

[1248] "Feedback" refers to data collected from the results of actual advertising placements (e.g., click-through rate, conversion rate) and used to improve the performance of the model.

[1249] The present invention relates to a system for recommending advertisements, which trains a machine learning model based on past data and current trend data, predicts future popularity, and recommends advertisement placements. Specific embodiments for implementing this system are described below.

[1250] System configuration and technologies used

[1251] The server sends requests to the APIs of social media platforms and news services (e.g., Twitter API, Google News API) to obtain current trending data, which is returned in JSON format and parsed by the server to extract information about specific topics or events.

[1252] The server uses the extracted keywords to collect past data from a database or another API (e.g., sports database API, Google Trends API). The collected data includes viewership, number of participants, number of social media posts, number of related news articles, etc. The server preprocesses this data and extracts features and target variables. Specifically, it imputes missing values ​​and normalizes the data.

[1253] The server uses the preprocessed data to train a model using a machine learning algorithm (e.g., linear regression model or random forest). This training process uses the Scikit-learn library, with features as input and popularity scores as output. The dataset is split into training data and test data, and the performance of the model is evaluated.

[1254] The server uses the trained model to predict future popularity based on current trend data. Specifically, it inputs data from the "Basketball World Cup 2023" obtained from the Twitter API into the model and calculates the predicted level of attention for the following weekend. The output is a popularity score, which is used to evaluate future trends.

[1255] The server recommends optimal advertising destinations based on the predicted popularity score. For example, if interest in the "Basketball World Cup 2023" is predicted to be at its highest next weekend, the server will recommend this trend as an advertising destination. It also provides specific advertising strategies and timing to advertisers.

[1256] The server collects the results data (e.g., click-through rate, conversion rate) after the ads are actually placed and uses them as feedback to improve the accuracy of the model. The results data are used to retrain the model, maximizing the effectiveness of advertising in the long term.

[1257] Examples of specific examples and prompts

[1258] For example, if the 2023 Basketball World Cup is trending rapidly on social media, the system would work like this:

[1259] 1. The server collects current trend data related to "BasketballWorldCup2023" from the Twitter API and Google News API, and checks the frequency of the extracted keyword "BasketballWorldCup2023."

[1260] 2. The server collects viewership data and participant numbers from the previous Basketball World Cup from the sports database API, completes the missing data, and extracts it as features.

[1261] 3. The server uses Scikit-learn to train a linear regression model based on past data and build a model to predict the popularity of events.

[1262] 4. The server uses the trained model to predict future interest in the Basketball World Cup 2023 and estimates the peak interest for the following weekend.

[1263] 5. Based on the prediction results, the server recommends "Basketball World Cup 2023" as an advertising destination and proposes specific advertising timing and strategies to the advertiser.

[1264] 6. The server collects post-advertising result data (click-through rate and conversion rate) and retrains the predictive model to improve accuracy.

[1265] Example prompt sentence:

[1266] "Collect current social media trends related to the 2023 Basketball World Cup, use past sporting event data to predict future popularity, and recommend optimal advertising destinations. Specifically, retrieve trend data from the Twitter API, train a linear regression model using Scikit-learn, and propose an advertising strategy based on the prediction results."

[1267] In this way, the system of the present invention provides advertisers with a means to effectively place advertisements in response to rapidly trending content in real time.

[1268] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1269] Step 1:

[1270] The server collects current trend data. It sends requests to APIs of social media platforms and news services (e.g., Twitter API, Google News API) to retrieve data about current topics. The input is the API request, and the output is trend data in JSON format. The server parses this data to extract key information about specific topics or events, such as the frequency of occurrence of the hashtag "BasketballWorldCup2023."

[1271] Step 2:

[1272] The server collects past data. Using the keywords extracted in step 1, the server retrieves past related event data from a database or another API (e.g., sports database API, Google Trends API). This includes viewership, number of participants, number of social media posts, number of related news articles, etc. The input is keywords related to past events, and the output is a dataset of related past events. The server parses this dataset to extract the necessary information.

[1273] Step 3:

[1274] The server preprocesses the data. The server normalizes the past data collected in step 2, fills in missing values, and formats it as features and target variables. For example, it formats viewer ratings and number of posts as numerical data. The input is raw data of past events, and the output is a preprocessed dataset.

[1275] Step 4:

[1276] The server trains the model. The server uses the preprocessed data to train the model using a machine learning algorithm (e.g., linear regression model or random forest). For example, it uses the Scikit-learn library. The input is the preprocessed dataset, and the output is the trained predictive model. The model is configured to input features and output a popularity score.

[1277] Step 5:

[1278] The server predicts future popularity. The server uses the trained model to predict future popularity from the current trend data obtained in step 1. The input is the current trend data and the trained model, and the output is the future popularity score. For example, the server predicts that the hashtag "BasketballWorldCup2023" will peak next weekend.

[1279] Step 6:

[1280] The server recommends advertising destinations. The server selects the optimal advertising destination based on the popularity score predicted in step 5 and recommends advertising to the advertiser. The input is the future popularity score, and the output is the recommended advertising destination. Specifically, the server recommends "Basketball World Cup 2023" as an advertising destination and suggests advertising strategies and timing.

[1281] Step 7:

[1282] The server collects feedback and improves the model. The server collects results after the ads are actually placed (e.g., click-through rate, conversion rate), and uses that data to retrain the model and improve its accuracy. The input is the results data after the ads are placed, and the output is an improved model. This makes it possible to maximize advertising effectiveness in the long term.

[1283] (Application example 1)

[1284] 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."

[1285] Conventional advertising systems statically select targets based on past data and are unable to respond to real-time trend changes, making it difficult to optimize advertising effectiveness. Furthermore, there is no feedback based on post-advertising results to improve model accuracy, meaning long-term improvements in advertising effectiveness cannot be expected. Furthermore, there is a lack of a way for users to easily evaluate advertising effectiveness and reflect that in their next ad placement.

[1286] 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.

[1287] In this invention, the server includes means for collecting past event data, means for training a model based on the collected past event data, means for collecting current trend data, means for predicting future popularity from the current trend data using the trained model, means for recommending advertising destinations based on the predicted popularity, means for users to provide feedback on the results of their advertising placements, and means for improving the accuracy of the model using the feedback results. This enables rapid response to trend fluctuations in real time and optimization of advertising effectiveness. Furthermore, improving the accuracy of the model based on user feedback can also be expected to improve long-term advertising effectiveness. Furthermore, users can easily evaluate advertising effectiveness and reflect it in their next advertising placements.

[1288] "Past event data" refers to information related to events that have taken place in the past, including viewership ratings, number of participants, number of posts on social media, number of related news articles, and the like.

[1289] "Training a model" is the process of using machine learning algorithms to learn relationships between collected data.

[1290] "Current trending data" is data on information and topics that are currently trending, collected from social media, news, and other sources.

[1291] "Predicting future popularity" is the process of using a trained model to infer what will be popular and trending in the future based on current trend data.

[1292] "Recommending advertising locations" means suggesting the most effective locations and timing for advertising based on predicted popularity.

[1293] "Providing feedback" means that the user reports to the server the results and effects obtained after placing an advertisement.

[1294] "Improving the accuracy of the model" means adjusting the parameters and algorithms of the predictive model based on the feedback results to increase the accuracy of the next prediction.

[1295] This invention relates to a system for recommending advertising, which consists of the following steps: First, a server collects past event data and trains a machine learning model based on this data. Next, the server collects current trend data and uses the trained model to predict future popularity. Based on the prediction results, the system recommends optimal advertising destinations to users.

[1296] Program processing

[1297] Data Collection Module

[1298] The server collects current trend data via social media and news APIs. Specifically, it uses Python to send requests to the Twitter API and Google News API, and retrieves JSON-formatted data including trending keywords and the number of posts.

[1299] Hardware used: Server, smartphone

[1300] Software used: Python, Twitter API, Google News API

[1301] Historical Data Analysis Module

[1302] The server collects past event data from a database or another API, and performs preprocessing to extract features and target variables. In this process, metrics such as viewership, number of participants, and number of posts on social media are used, and the data is organized using Python's Scikit-learn.

[1303] Hardware used: Server

[1304] Software used: Python, Scikit-learn

[1305] Model Training Module

[1306] The server uses the preprocessed data to train a machine learning algorithm (e.g., a linear regression model), which learns relationships in the data and generates a model to predict future popularity.

[1307] Hardware used: Server

[1308] Software used: Python, Scikit-learn

[1309] Popularity Prediction Module

[1310] The server uses the trained model to predict future popularity based on current trend data, a process also performed using Python.

[1311] Hardware used: Server

[1312] Software used: Python, Scikit-learn

[1313] Recommendation Engine

[1314] The server recommends the most effective advertising destinations based on predicted popularity, allowing users to know effective advertising destinations in real time.

[1315] Hardware used: Server

[1316] Software used: Python, Flask

[1317] Feedback Module

[1318] Users provide feedback on the results of their ad placements, and the data is sent to the server, where it is used to retrain the model, improving its prediction accuracy.

[1319] Hardware used: Smartphone

[1320] Software used: React Native, Firebase

[1321] Specific examples

[1322] As a specific application example, we will explain the process in which a server collects the latest trend data and predicts future popularity. For example, the server inputs the following prompt sentence into the generative AI model:

[1323] Gather the latest trends and predict future popularity. Use past event data to identify hot topics for the upcoming weekend and suggest them for advertising.

[1324] Through this process, the server can quickly respond to trend fluctuations in real time and optimize advertising effectiveness. It also has a feedback function to improve the accuracy of the model, making it a system that can achieve long-term improvements in advertising effectiveness.

[1325] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1326] Step 1:

[1327] The server collects past event data. Specifically, the server retrieves data related to past events, such as viewer ratings, number of participants, and number of posts on social media, from a database or API. After collection, this data is saved in JSON format and used for subsequent processing.

[1328] Input: Historical event data from a database or API

[1329] Output: Historical event data stored in JSON format

[1330] Step 2:

[1331] The server preprocesses the collected past event data and extracts features and target variables. Specifically, it performs missing value imputation, noise removal, and scaling of the data, and then extracts features (e.g., viewership rate, number of participants, number of social media posts) and target variables. Python's Scikit-learn is used.

[1332] Input: Historical event data stored in JSON format

[1333] Output: Preprocessed features and target variable

[1334] Step 3:

[1335] The server trains the model using the preprocessed data. Specifically, it uses a machine learning algorithm such as a linear regression model, inputs features, and outputs target variables. Python's Scikit-learn is used.

[1336] Input: Preprocessed features and target variables

[1337] Output: A trained machine learning model

[1338] Step 4:

[1339] The server collects current trend data from social media and news APIs. Specifically, it sends an API request and receives data in JSON format containing the latest topics and keywords. For example, it uses the Twitter API or Google News API.

[1340] Input: Trend data from Twitter API and Google News API

[1341] Output: Current trend data saved in JSON format

[1342] Step 5:

[1343] The server preprocesses the collected current trend data and extracts features. Specifically, it performs noise removal and normalization, and extracts features (e.g., topics, keywords, and number of posts). It uses Python's Scikit-learn.

[1344] Input: Current trend data stored in JSON format

[1345] Output: Preprocessed features

[1346] Step 6:

[1347] The server uses the trained model to predict future popularity based on current trend data. Specifically, the preprocessed features are input into the model to predict future popularity scores.

[1348] Input: Preprocessed features, trained model

[1349] Output: Predicted future popularity scores

[1350] Step 7:

[1351] The server recommends advertising destinations based on the predicted popularity. Specifically, it prioritizes trends with high popularity scores and suggests them as advertising destinations. Users receive this information through a smartphone application.

[1352] Input: Predicted future popularity scores

[1353] Output: Recommended ad placements

[1354] Step 8:

[1355] Users provide feedback on the results of their ads after placing them. Specifically, they use their smartphones to input advertising effectiveness measurement data (click rates, conversion rates, etc.) and send it to the server.

[1356] Input: Post-advertising effectiveness measurement data

[1357] Output: Feedback data sent to the server

[1358] Step 9:

[1359] The server uses the feedback data to improve the accuracy of the model. Specifically, it uses the feedback data as retraining data to adjust the model parameters and algorithms, thereby improving the accuracy of the next prediction.

[1360] Input: Feedback data

[1361] Output: A trained model with improved accuracy

[1362] 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.

[1363] The present invention combines an advertising recommendation system with an emotion engine that recognizes user emotions, and in addition to the process of collecting past event data, training a model based on this, collecting current trend data to predict future popularity, and recommending advertising destinations, the system utilizes user emotion data to achieve more accurate advertising placement. The program processing of this system is explained below in natural language.

[1364] Program Overview

[1365] The server retrieves the current trend

[1366] The server sends requests to social media and news APIs to retrieve current topics. This data, which contains information about specific topics or events, is returned in JSON format. The server analyzes this trend data to extract key keywords and temporal features.

[1367] The server collects and analyzes past event data

[1368] The server uses keywords extracted from current trends to collect past event data from a database or another API. This past data includes information about the popularity and attention of the event, such as viewer ratings, number of participants, number of social media posts, and number of related news articles. The server preprocesses this data and extracts features and target variables.

[1369] The server trains the model

[1370] The server uses the preprocessed historical event data to train a machine learning algorithm (e.g., a linear regression model). During this training process, features are input, popularity scores are output, and the model learns the relationship between them.

[1371] The server predicts future popularity

[1372] The server uses the trained model to predict future popularity from current trend data, allowing it to predict which events and content will attract attention in the future. The server adds the prediction results to the trend data and assigns a predicted popularity score to each trend.

[1373] The server recommends advertising

[1374] The server selects trends suitable for advertising based on the predicted popularity, prioritizes trends with high predicted popularity, and recommends the top trends that are expected to be particularly effective as advertising targets.

[1375] Introducing the Emotion Engine

[1376] The server uses an emotion engine to recognize the user's emotions. Specifically, it collects and analyzes the user's emotional data (e.g., comments and reactions posted on social media) to understand the user's emotional state in real time.

[1377] Analyzing Emotional Data

[1378] The server analyzes the collected emotion data and predicts the effectiveness of advertising. For example, it predicts that advertising will be more effective for topics that receive a lot of positive emotion, and prioritizes recommendations for those topics. Conversely, if there are a lot of negative emotions, it decides not to advertise on those topics.

[1379] Advertisement changes and adjustments based on emotions

[1380] The server can also recommend changes to ad content or creative based on user emotion recognition. For example, if a user's emotions are positive, it can suggest ads with positive messages. Conversely, if a user's emotions are predominantly negative, it can suggest ads with more neutral content.

[1381] Specific examples

[1382] For example, the following actions are taken for the 2023 Basketball World Cup:

[1383] 1. The server collects current topics from social media and news APIs:

[1384] The server sends a request to the API to retrieve trending data about the Basketball World Cup. For example, it detects that the hashtag "BasketballWorldCup2023" is popular.

[1385] 2. The server collects and analyzes past event data:

[1386] The server collects data on past sporting events (e.g., the last Basketball World Cup, the Olympics) and preprocesses their viewership and attendance figures.

[1387] 3. The server trains the model:

[1388] The server uses this historical data to train a machine learning model (e.g., a linear regression model).

[1389] 4. The server predicts future popularity:

[1390] The server uses the trained model to predict future popularity based on current trend data, and the prediction reveals that the Basketball World Cup 2023 will peak next weekend.

[1391] 5. The server recommends advertising:

[1392] Based on the prediction results, the server recommends "Basketball World Cup 2023" as an advertising destination. It assigns a score indicating the applicability and priority of the advertising destination and recommends the optimal timing for advertising.

[1393] 6. Introducing the Emotion Engine:

[1394] The server collects user reactions and comments from social media and analyzes them to understand user sentiment. For example, if the reaction to the "Basketball World Cup 2023" is positive, it determines that advertising will be effective.

[1395] 7. Advertisement modification and adjustment based on emotions:

[1396] If the server detects a user's positive emotions, it will suggest ads with positive messages. Conversely, if a user's negative emotions are detected, it will suggest more neutral ads.

[1397] In this way, the system of the present invention can not only respond to rapidly trending content in real time and effectively recommend advertising placements, but also take user emotions into consideration, thereby achieving even more effective advertising operations.

[1398] The processing flow will be explained below.

[1399] Step 1:

[1400] The server sends a request to a social media or news API to retrieve the most popular topics. Specifically, it makes a request to the Twitter API or Google News API to retrieve current trending data. The returned data is received in JSON format.

[1401] Step 2:

[1402] The server parses the API response and converts the trend data from JSON format into Python objects, making it easier to work with in your programs.

[1403] Step 3:

[1404] The server extracts keywords and temporal features (e.g., event dates and times, related hashtags) that are likely to be relevant to advertising from the trend data it acquires. This extraction is performed through text analysis of the trend data.

[1405] Step 4:

[1406] Based on the keywords extracted by the server, past event data is collected from a database or another API, and data on past sporting and entertainment events related to the extracted keywords is obtained.

[1407] Step 5:

[1408] The server preprocesses the collected historical event data. This preprocessing involves extracting features and target variables from data such as viewership, number of participants, and number of posts on social media. This prepares the data in a format suitable for model training.

[1409] Step 6:

[1410] The server uses the preprocessed data to train a machine learning algorithm (e.g., a linear regression model). Through the training process, the model learns the relationship between features and popularity scores.

[1411] Step 7:

[1412] The server uses the trained model to predict future popularity based on current trend data. Specifically, current trend data is input into the model, and predicted viewership ratings and social media reactions are output.

[1413] Step 8:

[1414] The server adds the predicted popularity score to the trend data, which gives each trend a numerical popularity score that can be used to determine advertising placements.

[1415] Step 9:

[1416] The server selects the best trend for advertising based on the popularity score. Specifically, it prioritizes trends with high popularity scores and lists them as recommended advertising destinations.

[1417] Step 10:

[1418] The server obtains user reactions and comments from the social media platform to collect user sentiment data, which includes the process of collecting reaction data through APIs.

[1419] Step 11:

[1420] The server analyzes the collected emotional data to identify the user's emotional state. Sentiment analysis uses natural language processing technology to classify emotions such as positive and negative from the content of comments.

[1421] Step 12:

[1422] The server changes or adjusts advertising destinations based on the emotion data. For example, it determines that topics with a high percentage of positive emotions have a high advertising effectiveness and recommends them as priority destinations.

[1423] Step 13:

[1424] The server combines emotion data with trend data to improve the accuracy of the model, enabling more accurate predictions that take into account the impact of emotional states on predictions.

[1425] Step 14:

[1426] The server notifies the user of recommended advertising placements. Specifically, it presents the user with a list of recommended advertising placements and provides detailed information on the optimal timing and location.

[1427] Step 15:

[1428] After the ad is placed, the server collects actual ad performance data, including viewer ratings, social media responses, and click counts.

[1429] Step 16:

[1430] The server analyzes the performance data collected and analyzes the difference between predicted and actual results, which allows the model to identify areas for improvement.

[1431] Step 17:

[1432] The server uses the feedback to retrain the model and improve its prediction accuracy. By incorporating real data, future predictions will be more accurate.

[1433] Through the above steps, the system of the present invention can respond to rapidly trending content in real time and realize effective advertising that takes into account user emotions.

[1434] Example 2

[1435] 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."

[1436] Conventional advertising systems typically place ads based on past data and current trend data. However, these systems do not take into account user emotional data, which limits the effectiveness of advertising and makes it difficult to optimally place ads. Particularly in today's world, where real-time data such as social media is increasingly being used, ignoring user emotions and reactions will result in a decrease in advertising effectiveness. Therefore, there is a demand for a system that takes user emotions into account and enables more accurate advertising placement.

[1437] 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 a means for collecting past event data, a means for training a model based on the collected past event data, and a means for collecting current trend data. This enables a means for collecting and analyzing user emotion data and a means for adjusting the content and creative of advertisements based on the emotion data.

[1438] "Past event data" refers to information such as viewership ratings, number of participants, number of social media posts, and number of related news articles related to a specific period or event.

[1439] "Training a model" refers to using collected data to train a machine learning algorithm to learn the relationship between features and target variables, making it usable for future predictions.

[1440] "Current trend data" refers to information about the latest topics and trends collected from social media and news APIs.

[1441] "Predicting future popularity" refers to using a trained model to calculate future attention and popularity from current trend data.

[1442] "Recommending advertising destinations" refers to suggesting optimal advertising investment destinations based on predicted popularity.

[1443] "User emotional data" refers to information on emotions and reactions obtained by analyzing comments and reactions posted by users on social media, etc.

[1444] "Collecting and analyzing emotional data" refers to collecting user comments and reactions and analyzing them to understand their emotional state, whether positive, negative, or neutral.

[1445] "Adjusting advertising content and creative based on emotional data" refers to optimizing advertising messages and visuals based on collected and analyzed emotional data.

[1446] This invention combines an advertising recommendation system with an emotion engine that recognizes user emotions. Specifically, it collects past event data, trains a model based on this, collects current trend data to predict future popularity, and recommends advertising destinations. In addition, by utilizing user emotion data, it achieves more accurate advertising placement.

[1447] The server sends requests to social media and news APIs (e.g., Twitter API, NewsAPI) to gather current topics. The data is often returned in JSON format, and the server parses it to extract keywords related to a particular topic or event. For example, the keyword "2023 Basketball World Cup" might be extracted.

[1448] Next, the server uses the keywords extracted from the current trends to collect past event data from a database or another API. This data includes detailed information about the popularity and attention of the event, such as viewership, number of participants, number of social media posts, and number of related news articles. The server preprocesses this data and extracts features and target variables.

[1449] The server then uses the preprocessed historical event data to train a machine learning algorithm (e.g., a linear regression model). Through this process, the model learns the relationship between features and popularity scores. Once trained, the model is used to predict future popularity from current trend data. This prediction process makes it possible to predict which events or content will attract attention in the future. The server adds the prediction results to the trend data and assigns a predicted popularity score to each trend.

[1450] Next, the server selects trends suitable for advertising based on the predicted popularity. For example, it prioritizes popular trends and recommends the top trends that are expected to be particularly effective as advertising targets.

[1451] Furthermore, the server will incorporate an emotion engine to collect and analyze user emotional data. Specifically, it will collect user reactions and comments from social media and analyze them to understand the user's emotional state in real time. For example, it will use IBM Watson's Natural Language Understanding API. This analysis will enable it to predict the effectiveness of advertising, predicting that advertising will be more effective for topics that attract a lot of positive emotions and prioritizing those topics for recommendations. Conversely, if there are a lot of negative emotions, it will decide not to advertise.

[1452] Furthermore, the server can recommend changes to the content and creative of advertisements based on the user's emotional state. For example, if the user's emotional state is positive, the server can suggest advertisements with positive messages. Conversely, if the user's emotional state is mostly negative, the server can suggest advertisements with more neutral content.

[1453] As a concrete example, for the 2023 Basketball World Cup, the following actions are taken:

[1454] The server uses the Twitter API to obtain trend data related to "BasketballWorldCup2023." For example, it detects that the hashtag "BasketballWorldCup2023" is frequently used. Next, the server collects and preprocesses viewership and attendance data related to previous Basketball World Cups and Olympics from Google Analytics. Next, the server trains a linear regression model using this historical data. Using the trained model, it predicts future popularity based on current trend data and predicts that "BasketballWorldCup2023" will peak next weekend. Based on the prediction results, the server then recommends "BasketballWorldCup2023" as an advertising destination. Next, the server uses IBM Watson's API to confirm positive sentiment toward "BasketballWorldCup2023" and suggests ads containing positive messages.

[1455] In this way, the system of the present invention can not only respond to rapidly trending content in real time and effectively recommend advertising placements, but also take user emotions into consideration, thereby achieving even more effective advertising operations.

[1456] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1457] Step 1:

[1458] The server retrieves the current trend.

[1459] The server sends requests to social media and news APIs (e.g., Twitter API, News API) to collect current topics. For example, to obtain trending data about the "Basketball World Cup 2023," the server makes the following API request:

[1460] Input: API request for "BasketballWorldCup2023"

[1461] Output: Trend data in JSON format

[1462] The server parses this JSON data and extracts keywords related to specific topics or events, using Python libraries to parse the JSON data and extract the necessary information.

[1463] Step 2:

[1464] The server collects and analyzes past event data.

[1465] The server uses keywords extracted from current trends and gathers data on past events from a database or another API, for example, data on past Basketball World Cups or Olympic Games.

[1466] Input: Current trending keyword (e.g. "Basketball World Cup")

[1467] Output: Past event data (e.g., viewership, number of participants, number of social media posts)

[1468] The server preprocesses this data and extracts features (e.g., viewership rate, number of SNS posts) and target variables (e.g., popularity score) using Python's pandas library.

[1469] Step 3:

[1470] The server trains the model.

[1471] The server uses the pre-processed historical event data to train a machine learning algorithm (e.g., a linear regression model).

[1472] Input: Preprocessed historical event data (features and target variables)

[1473] Output: The trained model

[1474] Implement and train a linear regression model using Python's scikit-learn library.

[1475] Step 4:

[1476] The server predicts future popularity.

[1477] The server uses the trained model to predict future popularity from current trend data.

[1478] Input: Features of the current trend data

[1479] Output: Predicted popularity score

[1480] The trained model is used to make predictions and assign each trend a predicted popularity score.

[1481] Step 5:

[1482] The server recommends advertising placements.

[1483] The server selects trends suitable for advertising based on the predicted popularity.

[1484] Input: Predicted popularity score

[1485] Output: A list of recommended ad placements

[1486] Based on the prediction results, the most popular trends are prioritized, and the top trends that are expected to be particularly effective are recommended as advertising targets.

[1487] Step 6:

[1488] The server implements the emotion engine.

[1489] The server uses an emotion engine to collect and analyze user emotion data, for example, by collecting user reactions and comments from social media and analyzing their emotions.

[1490] Input: Social media comments and reactions

[1491] Output: Parsed sentiment data (positive, negative, neutral, etc.)

[1492] Sentiment analysis is performed using IBM Watson's Natural Language Understanding API, among others.

[1493] Step 7:

[1494] The server analyzes the emotion data.

[1495] The server analyzes the collected emotional data and predicts the effectiveness of advertising placements.

[1496] Input: Parsed emotion data

[1497] Output: Prediction of advertising effectiveness

[1498] The system predicts that advertising will be effective for topics that have a lot of positive sentiment, and prioritizes recommendations for those topics. If there are a lot of negative sentiments, the system decides not to advertise on those topics.

[1499] Step 8:

[1500] The server modifies and adjusts the ads based on the sentiment.

[1501] The server recommends changes to the content and creative of the advertisement based on the user's emotion recognition.

[1502] Input: Parsed emotion data

[1503] Output: Proposal for changes to ad content

[1504] For example, if you have a lot of positive emotions, we'll suggest ads with positive messages, and if you have a lot of negative emotions, we'll suggest ads with more calming messages.

[1505] (Application example 2)

[1506] 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."

[1507] Previous ad placement recommendation systems focused on predicting ad placements based on past data and current trend data. However, ad placements that do not take user emotions and reactions into account may not achieve the expected results. Furthermore, by taking real-time reactions into account, it is necessary to flexibly adjust the content and timing of ads. Therefore, the challenge is to use user emotion data to achieve more accurate ad placements and maximize advertising effectiveness.

[1508] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting past data, means for training a model based on the collected past data, means for collecting current trend data, means for predicting future popularity from the current trend data using the trained model, means for recommending advertisement placements based on the predicted popularity, means for collecting and analyzing user emotion data, and means for adjusting the content of advertisements based on the analyzed emotion data. This enables highly accurate advertisement placement that takes user emotions into consideration.

[1509] "Past data" refers to data about events or occurrences that occurred in the past, including viewership ratings, number of participants, number of posts on social media, number of related articles, etc.

[1510] "Current trend data" is data about events and content that are currently trending in society and the market, and is obtained from social media and news articles.

[1511] "Future popularity" is a value that predicts how much attention and interest a topic will attract in the future based on current trend data.

[1512] A "model" is a mathematical structure that is trained by a machine learning algorithm to learn patterns from given data and make predictions or classifications.

[1513] "Recommending advertising locations" means proposing the optimal location and timing for placing an ad based on predicted future popularity.

[1514] "User emotional data" refers to data that indicates the emotions and reactions expressed by users on social media or other platforms, including positive and negative emotions.

[1515] An "emotion engine" is software that collects and analyzes a user's emotional data to recognize their emotions.

[1516] "Features" are characteristics or attributes used by models to learn data in data analysis, including things like an event's viewership rating or number of attendees.

[1517] A "target variable" is a variable that is the target of prediction by a prediction model, and in this invention, future popularity is the target variable.

[1518] "Preprocessing" refers to a series of steps taken to convert raw data into an analyzable format, including data cleaning and transformation.

[1519] "Analysis" is the process of extracting useful information and patterns from collected data.

[1520] "Adjusting the content of the advertisement" means changing the message or creative of the advertisement based on the user's emotional data.

[1521] The present invention provides a method for implementing a system that optimizes advertising placement using user emotion data, and an embodiment thereof will be described in detail below.

[1522] The server collects current trend data from social media and news APIs. For example, it retrieves how often a particular hashtag is used or which topics are mentioned in news articles. This data is returned in JSON format and analyzed. Key keywords and temporal features are extracted during analysis. This process is performed using Python's requests library and data analysis libraries (e.g., Pandas).

[1523] Next, the server collects past event data, including viewership, number of participants, number of social media posts, number of related news articles, etc. The collected data is preprocessed to extract features and target variables. This is done using a machine learning library (e.g., scikit-learn).

[1524] The server trains a machine learning model based on the preprocessed historical data. The training process uses the features as input and outputs a popularity score. The trained model is then applied to a linear regression model or other regression models.

[1525] The server also uses current trend data to predict future popularity, allowing it to predict which events and content will attract attention in the future. The predicted popularity is output in JSON format, and a score is assigned to each trend.

[1526] Furthermore, the server collects and analyzes users' emotional data. A sentiment analysis engine is used to analyze users' social media posts and reactions. This engine typically uses natural language processing libraries (e.g., Hugging Face transformers).

[1527] The server applies an algorithm to adjust the content of ads based on the emotional data. For example, ads with positive messages are delivered to users based on positive emotional data, while ads with more neutral content are recommended for users based on negative emotional data. Existing ad management systems (e.g., Google Ads) can be used to adjust ads.

[1528] A concrete example of server processing is the 2023 Basketball World Cup. The server collects current topics from social media and news APIs and detects that the hashtag "BasketballWorldCup2023" is frequently used. It collects data on past sporting events and preprocesses their viewership and attendance figures. It then uses a machine learning model to predict future popularity and predicts that "BasketballWorldCup2023" will peak next weekend. It also collects sentiment data from users' social media posts, and if there are a lot of positive reactions, it delivers ads with positive messages, thereby increasing advertising effectiveness.

[1529] Examples of prompts include:

[1530] Collect current trend data and use it to predict future popularity. Additionally, collect user reactions from social media and analyze sentiment. Based on the results, recommend the best advertising destinations.

[1531] As described above, the system of the present invention provides a technique for utilizing user emotion data to realize effective and highly accurate advertising.

[1532] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1533] Step 1:

[1534] The server collects current trend data from social media and news APIs. Specifically, it sends requests to the APIs and obtains the returned JSON-formatted data. The input is the endpoints of the social media and news APIs, and the output is JSON-formatted data about current topics. A data analysis library (e.g., Pandas) is used to extract key keywords and temporal features.

[1535] Step 2:

[1536] The server collects past event data. It retrieves event viewership, number of participants, number of social media posts, number of related news articles, etc. from a database or another API. The input is queries or API requests related to past events, and the output is past event data. This data is preprocessed to extract features and target variables. Preprocessing is performed using a machine learning library (e.g., scikit-learn).

[1537] Step 3:

[1538] The server trains a machine learning model based on preprocessed historical data. The input is the preprocessed features and target variable, and the output is the trained model. The model can be a linear regression model or another regression model. A machine learning library (e.g., scikit-learn) is used.

[1539] Step 4:

[1540] The server uses current trend data to predict future popularity. The input is the current trend data and a trained model, and the output is the future popularity score for each trend. The preprocessed current trend data is input into the machine learning model to perform predictions. The prediction results are output in JSON format.

[1541] Step 5:

[1542] The server collects and analyzes user emotion data. The input is user posts and reaction data from social media, and the output is the analyzed emotion data. An emotion analysis engine (e.g., Hugging Face transformers) is used to recognize positive, negative, and neutral emotions in real time.

[1543] Step 6:

[1544] The server recommends ad placements based on future popularity scores and user emotion data. The input is future popularity scores and analyzed emotion data, and the output is a list of optimal ad placements. Connected to the ad management system, the recommendation engine suggests optimal ad placement locations and timing.

[1545] Step 7:

[1546] The server adjusts the content of the advertisements based on the emotional data. The input is the analyzed emotional data, and the output is the adjusted advertising content. If the user's emotional data is positive, advertisements with positive messages are delivered, and if the emotional data is negative, advertisements with gentle content are recommended. This process uses an existing advertising management system (e.g., Google Ads Management).

[1547] The above are the specific processing steps for carrying out the present invention.

[1548] 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.

[1549] 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.

[1550] 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 robot 414.

[1551] 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.

[1552] 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.

[1553] 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.

[1554] 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).

[1555] 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.

[1556] 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."

[1557] 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.

[1558] 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).

[1559] 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.

[1560] 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.

[1561] 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.

[1562] 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.

[1563] 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.

[1564] 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.

[1565] 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.

[1566] 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.

[1567] 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.

[1568] 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.

[1569] The following is further disclosed regarding the above embodiment.

[1570] (Claim 1)

[1571] a means for collecting historical event data;

[1572] A means of training a model based on collected historical event data; and

[1573] a means of collecting current trend data;

[1574] A means for predicting future popularity from current trend data using the trained model;

[1575] A means for recommending advertising destinations based on the predicted popularity;

[1576] A system including:

[1577] (Claim 2)

[1578] A means to preprocess the collected historical event data and extract features and target variables;

[1579] means for training a model using the preprocessed data;

[1580] 10. The system of claim 1, further comprising:

[1581] (Claim 3)

[1582] A means for preprocessing the collected current trend data and extracting features;

[1583] A means of predicting future popularity based on pre-processed current trend data;

[1584] 10. The system of claim 1, further comprising:

[1585] "Example 1"

[1586] (Claim 1)

[1587] a means of collecting historical data;

[1588] A means of training a model based on collected historical data; and

[1589] a means of collecting current trend data;

[1590] A means for predicting future popularity from current trend data using the trained model;

[1591] a means for recommending advertising destinations based on the predicted popularity;

[1592] A means of obtaining and analyzing trend data from an API,

[1593] A means of preprocessing past data and extracting features and target variables;

[1594] a means for assigning a popularity score to each trend;

[1595] A means to collect feedback from actual ad placements and improve the model

[1596] A system including:

[1597] (Claim 2)

[1598] A means to preprocess the collected historical data and extract features and target variables;

[1599] means for training a model using the preprocessed data;

[1600] A means to show specific advertising strategies and timing,

[1601] 10. The system of claim 1, further comprising:

[1602] (Claim 3)

[1603] A means for preprocessing the collected current trend data and extracting features;

[1604] A means for predicting future popularity based on preprocessed current trend data;

[1605] A means to collect data on actual results and retrain the model,

[1606] a means for prioritizing placements according to the generated popularity scores;

[1607] 10. The system of claim 1, further comprising:

[1608] "Application Example 1"

[1609] (Claim 1)

[1610] a means for collecting historical event data;

[1611] A means of training a model based on collected historical event data; and

[1612] a means of collecting current trend data;

[1613] A means for predicting future popularity from current trend data using the trained model;

[1614] A means for recommending advertising destinations based on the predicted popularity;

[1615] A means for users to provide feedback on the results of their ads,

[1616] a means for improving the accuracy of the model using the feedback results;

[1617] A system including:

[1618] (Claim 2)

[1619] A means to preprocess the collected historical event data and extract features and target variables;

[1620] means for training a model using the preprocessed data;

[1621] A means for preprocessing the collected current trend data and extracting features;

[1622] 10. The system of claim 1, further comprising:

[1623] (Claim 3)

[1624] A means for recommending appropriate advertising destinations based on the prediction results;

[1625] A means for users to provide feedback on the results of their advertising,

[1626] a means for improving the accuracy of the model using feedback data; and

[1627] 10. The system of claim 1, further comprising:

[1628] "Example 2: Combining Emotion Engines"

[1629] (Claim 1)

[1630] a means for collecting historical event data;

[1631] A means of training a model based on collected historical event data; and

[1632] a means of collecting current trend data;

[1633] A means for predicting future popularity from current trend data using the trained model;

[1634] A means for recommending advertising destinations based on the predicted popularity;

[1635] A means for collecting and analyzing user emotion data;

[1636] A means to tailor advertising content and creative based on emotional data;

[1637] A system including:

[1638] (Claim 2)

[1639] A means to preprocess the collected historical event data and extract features and target variables;

[1640] means for training a model using the preprocessed data;

[1641] A means for preprocessing the collected current trend data and extracting features;

[1642] 10. The system of claim 1, further comprising:

[1643] (Claim 3)

[1644] A means for collecting and analyzing user emotional data in real time;

[1645] A means for predicting and adjusting the effectiveness of advertising based on emotion data;

[1646] 10. The system of claim 1, further comprising:

[1647] "Application example 2 when combining emotion engines"

[1648] (Claim 1)

[1649] a means of collecting historical data;

[1650] A means of training a model based on collected historical data; and

[1651] a means of collecting current trend data;

[1652] A means for predicting future popularity from current trend data using the trained model;

[1653] A means for recommending advertising destinations based on the predicted popularity;

[1654] means for collecting and analyzing user emotion data;

[1655] means for adjusting the content of the advertisement based on the analyzed emotion data;

[1656] A system including:

[1657] (Claim 2)

[1658] A means to preprocess the collected historical data and extract features and target variables;

[1659] means for training a model using the preprocessed data;

[1660] The system of claim 1 further comprising:

[1661] (Claim 3)

[1662] A means for preprocessing the collected current trend data and extracting features;

[1663] A means of predicting future popularity based on pre-processed current trend data;

[1664] The system of claim 1 further comprising: [Explanation of symbols]

[1665] 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 historical event data; A means of training a model based on collected historical event data; and a means of collecting current trend data; A means for predicting future popularity from current trend data using the trained model; A means for recommending advertising destinations based on the predicted popularity; A system including:

2. A means to preprocess the collected historical event data and extract features and target variables; means for training a model using the preprocessed data; The system of claim 1 further comprising:

3. A means for preprocessing the collected current trend data and extracting features; A means of predicting future popularity based on pre-processed current trend data; The system of claim 1 further comprising:

Images