System

Abstract

A system is provided.SOLUTION: A system comprising: means for collecting historical data for demand forecasting; means for pre-processing the collected data; means for training a machine learning model based on the pre-processed data; means for forecasting future demand using the trained machine learning model; means for providing forecast results to a user; and means for dynamically setting a price based on the demand forecast results.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 today's retail industry, demand forecasting errors and pricing difficulties are increasing, creating a need for more efficient inventory management and sales planning. Because a variety of factors, including market trends, seasonality, and customer behavior patterns, influence demand, there are limits to how much a retailer can predict manually. Therefore, a system that can solve these problems and achieve highly accurate demand forecasting and dynamic pricing is needed. [Means for solving the problem]

[0005] The present invention solves the problems by providing a system including: means for collecting historical data for demand forecasting; means for preprocessing the collected data; means for training a machine learning model based on the preprocessed data; means for forecasting future demand using the trained machine learning model; means for providing the forecast results to a user; and means for dynamically setting prices based on the demand forecast results. This system enables companies to optimize inventory management, streamline production and delivery schedules, and maximize profits through dynamic pricing.

[0006] "Demand forecasting" is a method of predicting future consumer demand by analyzing a variety of data.

[0007] "Historical data" refers to information such as previously recorded sales data and historical consumer behavior.

[0008] "Collection" is the act of gathering the necessary data from a specific location.

[0009] "Preprocessing" is a series of procedures that convert raw data into a form suitable for analysis and model training.

[0010] A "machine learning model" refers to the configuration of an algorithm that learns patterns from data and performs predictions, classifications, etc.

[0011] "Training" is the process of feeding data to a machine learning model and letting it learn.

[0012] "Pricing" is the act of determining the selling price of a product or service.

[0013] "Dynamic" refers to the state of continuously changing according to the situation or conditions.

[0014] "Providing" is the act of passing information or results to a user.

[0015] "System" refers to an entire device or program that is made up of multiple components with a set of functions.

[0016] "Inventory control" is the process of maintaining and managing the appropriate amount of product inventory.

[0017] "Production scheduling" is the process of systematically determining and managing the timing of product manufacturing.

[0018] "Delivery schedule" is the task of planning and managing the timing of product delivery. [Brief explanation of the drawings]

[0019] [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

[0020] 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.

[0021] First, the terms used in the following description will be explained.

[0022] 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).

[0023] 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.

[0024] 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.

[0025] 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.

[0026] 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."

[0027] [First embodiment]

[0028] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0029] 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.

[0030] 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).

[0031] 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.

[0032] 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.

[0033] 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.

[0034] 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.

[0035] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0036] 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.

[0037] 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.

[0038] 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.

[0039] 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."

[0040] This invention provides a system that uses AI and machine learning to predict future sales trends and demand. The specific processing of the program used in this system is shown below.

[0041] 1. Data Collection

[0042] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and other related systems via APIs.

[0043] Example: Last year's monthly sales data is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API.

[0044] 2. Data Preprocessing

[0045] The server imputes missing data from the collected raw data, for example, by imputing the average value or by connecting previous and next values.

[0046] The server detects erroneous or outlier values ​​and corrects or removes them as necessary.

[0047] The server converts the numeric data to a standard scale (e.g., a range from 0 to 1) to allow comparison between different data.

[0048] The server combines and formats the data collected from different sources into a single integrated data set.

[0049] 3. Model Training

[0050] The server splits the dataset into training data and test data, typically 80% as training data and 20% as test data.

[0051] The server selects and initializes a machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) suitable for demand forecasting.

[0052] The server uses the selected machine learning algorithm to build a model using the training data, for example, inputting the data into an LSTM network to train a demand forecasting model.

[0053] 4. Demand forecasting

[0054] The server inputs test data using the trained machine learning model to predict future sales trends and demand.

[0055] Example: Use test data to forecast sales for the next month and check the results.

[0056] 5. Outputting the results

[0057] The server formats the prediction results appropriately and sends them to the user's device.

[0058] Example: Sales forecast results are visualized as line graphs and bar graphs and displayed on the user's dashboard.

[0059] 6. Dynamic Pricing

[0060] A server runs an algorithm that dynamically sets prices for products based on demand forecasts.

[0061] For example, increase prices by 10% before a period when demand is predicted to increase, or lower prices during periods of excess inventory to encourage sales.

[0062] The server takes into account the current stock amount and performs processing to lower the price if there is excess stock and to raise the price if there is low stock.

[0063] The server reflects the optimized pricing in the product database and transmits the new price information to the user's terminal.

[0064] The system's forecasting and pricing capabilities enable users to highly optimize inventory management, production scheduling, and dynamic pricing, thereby enabling companies to make more effective decisions in sales planning and maximize profits.

[0065] The processing flow will be explained below.

[0066] Step 1:

[0067] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and other related systems via APIs. For example, monthly sales data from the last year is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API.

[0068] Step 2:

[0069] The server imputes missing data from the collected data, for example by imputing the average value or by using values ​​before and after the data, and then detects erroneous or outlier values ​​and corrects or deletes them as necessary.

[0070] Step 3:

[0071] The server converts the numerical data to a standard scale (e.g., a range from 0 to 1) to enable comparisons between different data sets, and combines and formats data collected from different sources into a single integrated dataset.

[0072] Step 4:

[0073] The server splits the formatted dataset into training data and test data, typically using 80% of the data as training data and 20% as test data.

[0074] Step 5:

[0075] The server selects and initializes the machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) that is best suited for demand forecasting. For example, it selects the LSTM model and sets the necessary hyperparameters.

[0076] Step 6:

[0077] The server trains the model using the selected machine learning algorithm by providing training data. Specifically, it inputs past sales data into an LSTM network and builds a model to predict future sales.

[0078] Step 7:

[0079] The server inputs test data using the trained model to predict future sales trends and demand, for example, forecasting sales for the next month, and outputs the results as data.

[0080] Step 8:

[0081] The server then converts the forecast results into an appropriate format and sends them to the user's device. For example, the sales forecast results are visualized as line graphs or bar charts and displayed on the user's dashboard.

[0082] Step 9:

[0083] The server runs an algorithm that dynamically sets product prices based on predicted demand, for example, raising prices by 10% ahead of periods of predicted increased demand and lowering prices during periods of overstock.

[0084] Step 10:

[0085] The server takes into account the current inventory level and optimizes pricing, for example lowering the price if inventory is excessive and raising the price if inventory is low.

[0086] Step 11:

[0087] The server reflects the optimized pricing in the product database and transmits the new price information to the user's terminal in real time.

[0088] Through this series of steps, the system of the present invention can achieve highly accurate demand forecasting and dynamic pricing, optimizing a company's inventory management and sales planning.

[0089] Example 1

[0090] 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."

[0091] In recent years, retailers have found it difficult to accurately forecast sales trends and demand, which has led to problems in inventory management and pricing. In particular, the processes of data collection, data preprocessing, model training, and dynamic pricing required to predict demand fluctuations are complex, and a system for managing these processes efficiently and effectively is needed.

[0092] 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.

[0093] In this invention, the server includes: means for collecting past data from retailers and related systems; means for preprocessing the collected data by imputing missing data, detecting and removing outliers, standard scale conversion, and data combination; means for dividing the preprocessed data into training data and test data and selecting and training a machine learning algorithm suitable for demand forecasting; means for predicting future sales trends and demand using the trained machine learning model; means for visually displaying the prediction results and sending them to a user's terminal; and means for dynamically setting product prices based on the demand forecast results and inventory levels and updating the product database. This enables accurate prediction of demand fluctuations and efficient and effective inventory management and pricing.

[0094] "Data collection" is the process of obtaining historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and related systems.

[0095] "Data preprocessing" is the process of filling in missing data from collected raw data, detecting and removing outliers, and converting the data into a standard scale to put the data into a unified format.

[0096] "Machine learning algorithms" are mathematical models or methods used to predict future trends or demand based on a dataset. This invention uses LSTM, Random Forest, and XGBoost as examples.

[0097] "Training data" refers to the large dataset used to build and train machine learning models.

[0098] "Test data" refers to a dataset used to evaluate the performance of a trained machine learning model.

[0099] "Demand forecasting" is the process of using trained machine learning models to predict future sales trends and product demand.

[0100] "Dynamic pricing" is the process of automatically adjusting product prices based on demand forecasts and inventory levels.

[0101] "Product database" means a digital database for managing and storing product and pricing information.

[0102] "Visual display" is the process of presenting the results of prediction to a user in a visually easy-to-understand format such as a graph or table.

[0103] "User terminal" refers to a device, such as a PC or smartphone, that receives and displays prediction results and price information.

[0104] This invention provides a system that uses AI and machine learning to predict future sales trends and demand. This system operates in cooperation with a server, terminals, and users.

[0105] Data collection

[0106] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and related systems via APIs, specifically retrieving sales data from an SQL database and using the Google Analytics API to pull web browsing history data.

[0107] Example: Retrieve last year's monthly sales data using an SQL query and collect user web browsing history using the Google Analytics API.

[0108] Data Preprocessing

[0109] The server performs preprocessing on the collected raw data, including missing data imputation, outlier detection and removal, standard scale conversion, and data merging. Missing data is imputed using mean imputation, and outliers are filtered based on standard deviation. Data is standardized to a range of 0 to 1, and different data sources are combined into an integrated dataset.

[0110] Example: Use the Pandas library to impute NaN values ​​in a DataFrame, remove outliers, and perform normalization.

[0111] Model learning

[0112] The server splits the preprocessed data into training data and test data, selects and initializes a machine learning algorithm such as LSTM, Random Forest, or XGBoost, and builds and trains a model based on the training data.

[0113] Example: Using Sci-kit Learn, split the dataset into 80% training data and 20% test data, and build and train an LSTM model.

[0114] Demand forecasting

[0115] The server uses the trained model to predict future sales trends and demand based on the test data, and the prediction results are saved in a format such as a CSV file.

[0116] Example: Use test data to forecast sales for the next month and save the results to a file.

[0117] Output of results

[0118] The server visually displays the prediction results and sends them to the user's device. For example, it uses Matplotlib or Seaborn to draw line graphs and bar charts and display them on a dashboard.

[0119] Example: Generate a graph based on the prediction results and display it on a dashboard.

[0120] Dynamic Pricing

[0121] The server dynamically sets product prices based on demand forecasts and inventory levels, and updates the product database with this information. Based on a pricing algorithm, prices are increased during periods of increased demand and decreased when inventory levels are excessive.

[0122] Example: Raise prices by 10% when demand is predicted to increase, or lower prices when there is excess inventory.

[0123] Example prompt: "Use AI and machine learning to generate a Python script that uses the past year's sales data to forecast next month's sales and dynamically set pricing. Please include any libraries needed and specific code examples."

[0124] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0125] Step 1: Data collection

[0126] The server collects historical sales data, consumer behavior data, seasonal data, and market trend data from retailers and related systems. It queries an SQL database and retrieves web browsing history data using the Google Analytics API. The input is sales data, web click data, etc., and the output is a consolidated dataset containing these data. Specifically, the server securely collects data using API keys and authentication tokens.

[0127] Step 2: Data Preprocessing

[0128] The server preprocesses the collected data, including imputing missing data, detecting and removing outliers, converting to a standard scale, and combining different data sources. The input is raw data, and the output is preprocessed, clean data. The server uses the Pandas library to process the missing data, such as "df['Sales'].fillna(df['Sales'].mean(), inplace=True)", to filter outliers and convert to a standard scale. It also combines data from different data sources to generate a unified dataset.

[0129] Step 3: Split the data

[0130] The server splits the preprocessed data into training data and test data. The input is the clean dataset, and the output is the training data and test data. The server uses Sci-kit Learn and executes code such as "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)".

[0131] Step 4: Model training

[0132] The server uses the training data to train a machine learning model. The input is the training data, and the output is the trained model. The server selects an algorithm such as LSTM, Random Forest, or XGBoost, initializes it with code such as "from keras.models import Sequential" and "model = Sequential()", and trains the model with code such as "model.fit(X_train, y_train, epochs=50, batch_size=32, validation_split=0.2)".

[0133] Step 5: Demand forecast

[0134] The server uses the trained model to predict future sales trends and demand based on the test data. The input is the test data and the output is the prediction result. The server makes the prediction using "y_pred = model.predict(X_test)" and saves the prediction result in a format such as a CSV file.

[0135] Step 6: Output the results

[0136] The server visually displays the prediction results and sends them to the user's device. The input is the prediction results, and the output is the visually displayed data. The server uses Matplotlib and Seaborn to draw graphs, visualizes them with code such as "plt.plot(prediction result)", and displays them on a dashboard.

[0137] Step 7: Dynamic Pricing

[0138] The server dynamically sets product prices based on demand forecast results and inventory levels, and reflects this information in the product database. The inputs are forecast results and inventory data, and the output is dynamically set price information. The server adjusts prices during periods of increased demand, such as "item_price = 1.10", and updates the new price information in the product database. Conditional branching such as "if inventory level > threshold: price = 0.90" and "elif inventory level < threshold: price = 1.10" is used. Specifically, the server updates price information via an SQL query and sends the new price information to the user's device.

[0139] (Application example 1)

[0140] 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."

[0141] Conventional demand forecasting systems require a lot of time and resources to compile data and make forecasts, making it difficult to respond quickly to fluctuations in demand. Furthermore, there are limited ways to present forecast results in a visually understandable manner, making it difficult for managers and store managers to make appropriate decisions. Furthermore, when it comes to dynamic pricing, there is a lack of functionality to automatically present effective strategies based on demand forecasts.

[0142] 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.

[0143] In this invention, the server includes means for collecting past data for demand forecasting, means for preprocessing the collected data, means for training a machine learning model based on the preprocessed data, means for forecasting future demand using the trained machine learning model, means for providing the forecast results to a user, means for dynamically setting prices based on the demand forecast results, means for predicting sales trends and proposing appropriate price increases or decreases, and means for visualizing the forecast results, thereby enabling store managers to quickly and effectively perform demand forecasting and optimize dynamic pricing and inventory management.

[0144] "Demand forecasting" is the prediction of future customer demand and sales volume based on past data.

[0145] "Data collection" refers to the collection of historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and related systems.

[0146] "Data preprocessing" refers to processing collected raw data, such as filling in missing data, correcting or deleting error values ​​and outliers, converting numerical data to a standard scale, and integrating different data.

[0147] A "machine learning model" is an algorithm that is trained to perform tasks such as prediction or classification using collected and preprocessed data as input.

[0148] "Dynamic pricing" refers to adjusting the price of goods and services in real time in response to changes in supply and demand.

[0149] "Collected Data" refers to historical sales data, consumer behavior data, seasonal data, and market trend data obtained from retailers and related systems via APIs or other means.

[0150] A "machine learning algorithm" is a method or procedure for performing machine learning tasks such as demand forecasting and classification based on various data, for example, LSTM, Random Forest, or XGBoost.

[0151] "Sales trend forecasting" refers to predicting future sales figures and sales trends based on past sales data and other related information.

[0152] A "price increase or decrease proposal" is a specific recommendation to increase or decrease the price of a product based on projected demand and inventory availability.

[0153] "Visualization of forecast results" refers to displaying predicted demand and sales trends in a visual format such as graphs and charts.

[0154] This invention relates to a demand forecasting and dynamic pricing system that utilizes AI and machine learning to predict future demand based on a wide range of data, including past sales data, and dynamically sets prices for products and services based on the results.

[0155] First, the server collects past sales data, consumer behavior data, seasonal data, and market trend data from retailers and related systems via API. For example, monthly sales data from the last fiscal year is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API. The data collected in this way is stored on the server as raw data.

[0156] Next, the server preprocesses the collected data. Missing data is imputed with the average value, and errors and outliers are detected and corrected or removed. Collected data from different data sources is integrated, converted to a standard scale (e.g., a range from 0 to 1), and organized into a unified dataset. The main software used in this process includes Python's Pandas and Numpy.

[0157] The server trains a machine learning model based on the preprocessed data. The dataset is divided into training data and test data, and a machine learning algorithm such as random forest or LSTM is selected to build the model. For example, past sales data and consumer behavior data are input to train the LSTM network. Libraries such as Scikit-learn and TensorFlow are used in this process.

[0158] The server uses the trained machine learning model to predict future demand. Test data can be input into the model to predict future sales trends. The prediction results are sent to the user's device and displayed in a visualized format, such as a line graph or bar chart. The results are presented in an intuitive and easy-to-understand format for managers and store administrators.

[0159] The server also has a means for dynamically setting prices based on the results of demand forecasts. It automatically adjusts prices by raising them during periods of increased demand and lowering them when there is excess inventory. Prices are dynamically set in conjunction with current inventory levels, promoting sales. This algorithm is reflected in the product database, and updated price information is sent to users' devices in real time.

[0160] For example, when weekly sales data, consumer behavior data, and market trend data are input, the system will suggest appropriate price increases or decreases based on the predicted sales trends for the following month.An example of a prompt to be input into the generative AI model is, "Based on last week's sales data, consumer behavior data, and market trend data, please suggest predicted sales trends and dynamic pricing for the next two weeks."

[0161] In this way, the present invention provides an efficient tool for predicting future sales trends and dynamic pricing in brick-and-mortar stores.

[0162] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0163] Step 1: Data collection

[0164] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and related systems via APIs. This may include, for example, retrieving monthly sales data from an SQL database or collecting customer web browsing history through the Google Analytics API. The input of this step is the data source, and the output is the collected raw data.

[0165] Step 2: Data Preprocessing

[0166] The server preprocesses the collected raw data by first imputing missing data with the mean value, then detecting and correcting or removing erroneous values ​​and outliers, normalizing numerical data to a range of 0 to 1, and integrating data collected from different data sources. The input of this step is the collected raw data, and the output is a preprocessed dataset.

[0167] Step 3: Train the model

[0168] The server uses the preprocessed data to train a machine learning model. First, it splits the dataset into training data and test data. Then it selects an algorithm, such as random forest or LSTM, and builds a model using the training data. The input of this step is the preprocessed data, and the output is a trained machine learning model.

[0169] Step 4: Demand forecast

[0170] The server uses the trained machine learning model to predict future demand. Specifically, it inputs test data into the model and calculates future sales trends. The inputs for this step are the test data and the trained model, and the output is predicted demand data.

[0171] Step 5: Providing prediction results

[0172] The server provides the forecast results to the user. For example, it displays the forecasted demand data on the user's terminal in a visual format such as a line graph or bar graph. The input of this step is the forecasted demand data, and the output is the visualized forecast results.

[0173] Step 6: Dynamic Pricing

[0174] The server dynamically sets prices based on the demand forecast results. Specifically, it raises product prices before the period when demand is predicted to increase, and lowers prices when there is excess inventory. It also takes into account current inventory levels and raises prices when inventory is low. The inputs to this step are predicted demand data and inventory data, and the output is dynamically set price information.

[0175] Step 7: Sales trend forecast and proposals

[0176] The server proposes price increases or decreases based on future sales trends. For example, by inputting weekly sales data, consumer behavior data, and market trend data, the server proposes appropriate price adjustments based on the predicted sales trends for the next month. The inputs for this step are the collected data and predicted sales trends, and the output is a specific price adjustment proposal.

[0177] Step 8: Visualize the prediction results

[0178] The server visualizes the predicted demand and sales trends. It displays the forecast results in line graphs and bar charts, and communicates them in an intuitive and easy-to-understand format to the user. The input to this step is the predicted demand data and sales trend data, and the output is visualized graphs and charts.

[0179] 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.

[0180] This invention combines a system that uses AI and machine learning to predict future sales trends and demand with an emotion engine that recognizes user emotions. The specific processing of the program used in this system is shown below.

[0181] 1. Data Collection

[0182] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and other related systems via APIs. For example, monthly sales data from the last year is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API.

[0183] 2. Data Preprocessing

[0184] The server imputes missing data from the collected data, for example by imputing the average value or by using values ​​before and after the data, and then detects erroneous or outlier values ​​and corrects or deletes them as necessary.

[0185] The server converts the numerical data to a standard scale (e.g., a range from 0 to 1) to enable comparisons between different data sets, and combines and formats data collected from different sources into a single integrated dataset.

[0186] 3. Model Training

[0187] The server splits the formatted dataset into training data and test data, typically using 80% of the data as training data and 20% as test data.

[0188] The server selects and initializes the machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) that is best suited for demand forecasting. For example, it selects the LSTM model and sets the necessary hyperparameters.

[0189] The server trains the model using the selected machine learning algorithm by providing training data. Specifically, it inputs past sales data into an LSTM network and builds a model to predict future sales.

[0190] 4. Demand forecasting

[0191] The server inputs test data using the trained model to predict future sales trends and demand, for example, forecasting sales for the next month, and outputs the results as data.

[0192] 5. Outputting the results

[0193] The server then converts the forecast results into an appropriate format and sends them to the user's device. For example, the sales forecast results are visualized as line graphs or bar charts and displayed on the user's dashboard.

[0194] The emotion engine uses a camera and microphone to analyze the user's facial expressions and voice when presenting results, and dynamically changes the way prediction results are presented based on the recognized emotions.

[0195] For example: If the user is surprised, show them some additional information to help them calm down. If the user is confused, show them a detailed explanation.

[0196] 6. Dynamic Pricing

[0197] The server runs an algorithm that dynamically sets product prices based on predicted demand, for example, raising prices by 10% ahead of periods of predicted increased demand and lowering prices during periods of overstock.

[0198] The server takes into account the current inventory level and optimizes pricing, for example lowering the price if inventory is excessive and raising the price if inventory is low.

[0199] The server reflects the optimized pricing in the product database and transmits the new price information to the user's terminal in real time.

[0200] In addition to the demand forecasting and dynamic pricing functions provided by this system, by integrating user emotion recognition with an emotion engine, it is possible to provide a more intuitive and adaptive interface for users, which will enable companies to improve not only inventory management and production scheduling but also the user experience.

[0201] The processing flow will be explained below.

[0202] Step 1:

[0203] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and other related systems via APIs. For example, monthly sales data from the last year is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API.

[0204] Step 2:

[0205] The server imputes missing data from the collected data, for example by imputing the average value or by using values ​​before and after the data, and then detects erroneous or outlier values ​​and corrects or deletes them as necessary.

[0206] Step 3:

[0207] The server converts numerical data to a standard scale (e.g., a range from 0 to 1) to allow comparisons between different data sets, and combines and formats data collected from different sources into a single integrated dataset.

[0208] Step 4:

[0209] The server splits the formatted dataset into training data and test data, typically using 80% of the data as training data and 20% as test data.

[0210] Step 5:

[0211] The server selects and initializes the machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) that is best suited for demand forecasting. For example, it selects the LSTM model and sets the necessary hyperparameters.

[0212] Step 6:

[0213] The server trains the model using the selected machine learning algorithm by providing training data. Specifically, it inputs past sales data into an LSTM network and builds a model to predict future sales.

[0214] Step 7:

[0215] The server inputs test data using the trained model to predict future sales trends and demand, for example, forecasting sales for the next month, and outputs the results as data.

[0216] Step 8:

[0217] The server then converts the forecast results into an appropriate format and sends them to the user's device. For example, the sales forecast results are visualized as line graphs or bar charts and displayed on the user's dashboard.

[0218] Step 9:

[0219] The device uses a camera and microphone to collect the user's facial expressions and voice data, which the emotion engine analyzes to recognize the user's emotions, such as whether the user is confused or surprised.

[0220] Step 10:

[0221] The server dynamically changes how the predictions are presented based on the recognized emotion: for example, if the user is surprised, it may display additional information to calm them down, or if the user is confused, it may display additional detailed explanations or instructions.

[0222] Step 11:

[0223] The server runs an algorithm that dynamically sets product prices based on predicted demand, for example, raising prices by 10% ahead of periods of predicted increased demand and lowering prices during periods of overstock.

[0224] Step 12:

[0225] The server takes into account the current inventory level and optimizes pricing, for example lowering the price if inventory is excessive and raising the price if inventory is low.

[0226] Step 13:

[0227] The server reflects the optimized pricing in the product database and transmits the new price information to the user's terminal in real time.

[0228] Step 14:

[0229] Users can check new price information and forecast results through their devices, and can adjust their inventory management and sales strategies accordingly.

[0230] Through this series of steps, the system of the present invention achieves highly accurate demand forecasting and dynamic pricing, and also provides an adaptive interface that responds to user emotions, allowing companies to optimize all aspects of inventory management, product production, and sales strategies.

[0231] Example 2

[0232] 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."

[0233] Conventional demand forecasting systems use past data to predict future demand, but they are unable to provide an interface that takes user emotions into account, resulting in a poor user experience. In particular, because the user's psychological state is not reflected when presenting the forecast results, there is a lack of support for the user to understand the results and take action. Conventional systems also lack the ability to dynamically set prices.

[0234] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0235] In this invention, the server includes means for collecting past data for demand forecasting, means for preprocessing the collected data, means for training a machine learning model based on the preprocessed data, means for forecasting future demand using the trained machine learning model, means for providing the forecast results to a user, means for recognizing the user's emotions, means for dynamically presenting the forecast results based on the user's emotions, and means for dynamically setting prices based on the demand forecast results. This makes it possible to provide an interface that reflects the user's emotions, improving the user experience and providing more support for understanding and taking action on the demand forecast results. Furthermore, dynamic pricing enables flexible price management in response to market fluctuations.

[0236] "Demand forecasting" is the process of predicting future consumer demand based on past and current data.

[0237] "Data Collection Measures" refers to the ability to collect necessary data from retailers and other related systems.

[0238] "Data preprocessing means" refers to functions that fill in missing values ​​in collected data, correct or delete error values ​​or outliers, and perform scaling and integration.

[0239] "Machine learning model training means" refers to a function that uses preprocessed data to train a machine learning algorithm and create a predictive model.

[0240] "Demand forecasting means" refers to the ability to forecast future demand using a trained machine learning model.

[0241] "Means of providing to users" refers to the function of providing prediction results to users in an appropriate format and presenting them in an easy-to-understand manner.

[0242] "Emotion recognition means" refers to a function that analyzes the user's facial expressions and voice to recognize the user's emotional state.

[0243] "Dynamic presentation means" refers to a function that dynamically changes the way prediction results are presented based on the recognized user emotions.

[0244] "Dynamic pricing means" refers to a function that dynamically sets product prices in real time based on demand forecast results.

[0245] This invention combines a system that uses AI and machine learning to predict future sales trends and demand with an emotion engine that recognizes user emotions. The specific processing of the program used in this system is described in detail below.

[0246] First, the server collects data. It collects past sales data, consumer behavior data, seasonal data, market trend data, etc. from retailers and related systems. Specifically, it retrieves last year's monthly sales data from an SQL database and collects users' web browsing history using the Google Analytics API. The data collected in this way includes information from a variety of sources.

[0247] Next, the server performs data preprocessing. Missing data is imputed from the collected data, and errors and outliers are detected and corrected or removed. Specific processing methods include average imputation and imputation based on surrounding values. All numerical data is also converted to a standard scale to enable comparisons between different data sets. Data collected from different sources is combined into a single integrated dataset and formatted.

[0248] The server then performs model training. The formatted dataset is divided into training data and test data, and the machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) optimal for demand forecasting is selected and initialized. For example, an LSTM model is selected and the necessary hyperparameters are set. Training data is input using this selected algorithm to train the model. Specifically, past sales data is input into the LSTM network to build a model for predicting future sales.

[0249] Next, the server performs demand forecasting. It uses the trained model to input test data and predict future sales trends and demand. For example, it predicts sales for the next month and outputs the results as data.

[0250] The prediction results are provided from the server to the user's device in an appropriate format. The prediction results are visualized in line graphs and bar charts and displayed on the user's dashboard. The emotion engine uses a camera and microphone to analyze the user's facial expressions and voice when presenting the results, recognizing the user's emotions. Based on the recognized emotions, the way the prediction results are presented is dynamically changed. For example, if the user is surprised, additional information to calm them down is displayed, and if the user is confused, a detailed explanation is displayed.

[0251] Dynamic pricing is also an important feature of this system. The server runs an algorithm that dynamically sets product prices based on predicted demand. For example, the price can be raised by 10% before a period when demand is predicted to increase, and lowered when there is excess inventory. Furthermore, pricing is optimized taking into account current inventory levels. For example, if there is excess inventory, the price is lowered, and if inventory is low, the price is raised. The optimized pricing is reflected in the product database, and new price information is sent to users' devices in real time.

[0252] The system integrates demand forecasting and dynamic pricing functions, as well as user emotion recognition through an emotion engine, to provide a more intuitive and adaptive interface for users, helping companies not only to optimize inventory management and production schedules, but also to improve the user experience.

[0253] Specific examples

[0254] For example, the server retrieves last year's sales data from an SQL database and collects users' web browsing history using the Google Analytics API. It then uses mean imputation to impute missing data and scales the data to a range of 0 to 1 using MinMaxScaler. It selects an LSTM model and trains it using the training data with the fit() function. It uses the predict() function to predict next month's sales and displays the results in a line graph. The emotion engine analyzes emotions from the user's facial expressions and voice and dynamically changes the way the prediction results are presented. Based on demand, the server may raise prices by 10% or optimize prices, update the product database, and notify the user's device of the new price information.

[0255] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0256] Program processing flow and specific explanation

[0257] Step 1: Data collection

[0258] Input: Database connection information for retailers and related systems, API keys

[0259] Specific behavior:

[0260] The server uses an SQL query to retrieve last year's monthly sales data from the database.

[0261] The server uses the Google Analytics API to collect the user's web browsing history for the past six months.

[0262] Output: Collected dataset of historical sales data and consumer behavior data

[0263] Step 2: Data Preprocessing

[0264] Input: raw data collected, missing data, data containing error values ​​and outliers

[0265] Specific behavior:

[0266] The server generates a data frame using the Pandas library and performs imputation of missing values ​​(e.g., mean imputation).

[0267] The server runs an outlier detection algorithm to detect and remove or correct outliers.

[0268] The server applies MinMaxScaler to scale the numeric data to the range 0 to 1.

[0269] The server combines and formats the data collected from different sources into a single integrated data set.

[0270] Output: Preprocessed consolidated dataset

[0271] Step 3: Model training

[0272] Input: Preprocessed unified dataset, dataset split into training and test data

[0273] Specific behavior:

[0274] The server splits the combined dataset into training data (80%) and test data (20%) using the Scikit-learn or Keras libraries.

[0275] The server selects an LSTM model and sets hyperparameters (e.g., learning rate, number of epochs).

[0276] The server uses the training data to train the LSTM model using the fit() function.

[0277] Output: A trained LSTM model

[0278] Step 4: Demand forecast

[0279] Input: trained model, test data

[0280] Specific behavior:

[0281] The server uses the predict() function to forecast future sales trends and demand based on the test data.

[0282] Specifically, it generates sales forecast results for the next month.

[0283] Output: Future sales forecast results

[0284] Step 5: Output the results

[0285] Input: Future sales forecast results

[0286] Specific behavior:

[0287] The server uses the Matplotlib and Seaborn libraries to visualize the forecast results in line and bar graphs.

[0288] The server sends the visualized data to the user's device and displays it on a dashboard.

[0289] Output: Visualized data for dashboard display

[0290] Step 6: Emotion Recognition

[0291] Input: User's facial expressions, voice data

[0292] Specific behavior:

[0293] The emotion engine analyzes facial and voice data collected from the camera and microphone to recognize the user's emotional state.

[0294] The emotion engine uses facial recognition and voice analysis algorithms.

[0295] Output: Recognized emotional state of the user

[0296] Step 7: Dynamic Presentation

[0297] Input: Recognized user emotional state, future sales forecast results

[0298] Specific behavior:

[0299] The server dynamically changes the way predictions are presented based on the user's perceived emotions: if the user is surprised, it displays additional reassuring information, and if the user is confused, it displays a detailed explanation.

[0300] Output: Dynamically changed presentation results based on user emotion

[0301] Step 8: Dynamic Pricing

[0302] Input: Future sales forecast results, current inventory amount

[0303] Specific behavior:

[0304] The server dynamically sets product prices based on predicted demand, for example by raising prices by 10% before periods when demand is predicted to increase, and lowering prices when there is excess inventory.

[0305] The server optimizes pricing by taking into account inventory levels. Specifically, if inventory is excessive, the price is reduced by 5%, and if inventory is low, the price is increased by 5%.

[0306] Output: Optimized pricing information

[0307] Step 9: Update and submit pricing information

[0308] Input: Optimized Pricing Information

[0309] Specific behavior:

[0310] The server updates the product database and transmits new price information to the user terminal in real time.

[0311] Output: Updated price information, real-time notification to user terminal

[0312] (Application example 2)

[0313] 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."

[0314] Conventional demand forecasting systems primarily use historical data and machine learning to predict future sales trends, but they are unable to consider the user's emotional state. As a result, it is difficult to provide optimal interfaces and adaptive feedback for users. In particular, the lack of emotion recognition makes it difficult for users to accept and understand the prediction results, resulting in reduced usability. Furthermore, while dynamic pricing is automatically set based on demand forecasts, it fails to sufficiently increase user satisfaction because pricing based on user emotions and behavior is not taken into account.

[0315] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0316] In this invention, the server includes means for collecting past data for demand forecasting, means for preprocessing the collected data, means for training a machine learning model based on the preprocessed data, means for forecasting future demand using the trained machine learning model, means for providing the forecast results to a user, means for dynamically setting prices based on the demand forecast results, and means for recognizing the user's emotions and dynamically changing the presentation method of the forecast results and the interface based on the emotions. This makes it possible to provide more intuitive and adaptive feedback while taking the user's emotions into consideration. Dynamic pricing based on the user's emotions is also realized, thereby improving user satisfaction.

[0317] "Demand forecasting" refers to analyzing past data and predicting future sales trends and demand.

[0318] "Means for collection" refers to a function that allows necessary data to be acquired from various data sources.

[0319] "Means for data preprocessing" refers to the function of completing missing data and correcting outliers in collected data, and processing it to prepare it in a format that can be analyzed.

[0320] A "machine learning model" is a trained algorithm used to predict future demand based on historical data.

[0321] "Training" refers to the process of inputting data into a machine learning model and adjusting the model's parameters to improve its predictive accuracy.

[0322] A "demand forecasting means" is one that has the ability to forecast future sales trends and demand using trained machine learning models.

[0323] "Means for providing to the user" refers to a function that provides an interface for displaying prediction results to the user.

[0324] "Dynamic pricing" refers to the ability to adjust product prices in real time based on predicted demand.

[0325] "Means for recognizing user emotions" refers to technology that uses a camera and microphone to analyze the user's facial expressions and voice to identify their emotional state.

[0326] "Means for dynamically changing the interface presentation method" refers to a function that adaptively changes the screen display and information presentation method based on the user's emotions.

[0327] A system for implementing the present invention is configured as follows.

[0328] First, the server collects historical data for demand forecasting. The collected data includes past sales data, consumer behavior data, seasonal data, and market trend data. This data is collected by querying an SQL database or retrieving data through APIs.

[0329] The server then preprocesses the collected data, which includes imputing missing data, correcting errors and outliers, converting numerical data to a standard scale, and integrating data from different sources. This process prepares the data in a format that can be analyzed.

[0330] The server then trains a machine learning model based on the preprocessed data. The machine learning model to be used may be LSTM, Random Forest, XGBoost, or other models. The data is separated into training data and test data, and the model is trained using the selected algorithm. This process builds a highly accurate model for predicting future demand.

[0331] The server uses the trained model to predict future demand, and the predicted demand data is provided to the user. For example, the sales forecast results are displayed on the user's device as a line graph or bar chart.

[0332] The server also uses a camera and microphone to recognize the user's emotions. Using audio and image analysis software such as EmotionEngine, the server analyzes the user's facial expressions and voice to identify their emotions. Based on this identified emotion, the server dynamically changes the way prediction results are presented and the interface. For example, if the user is surprised, the server displays additional information along with the prediction result.

[0333] The server also dynamically sets prices based on the results of demand forecasts. Prices are raised if demand is predicted to increase, and lowered if there is excess inventory. This allows companies to optimize inventory management and pricing.

[0334] This system provides the following advantages: Improved usability, as users can intuitively understand the information they want; and Dynamic pricing, which maximizes profits for businesses. This is a specific embodiment of the present invention.

[0335] For example, if the user is surprised, a message will appear on the screen saying, "Don't worry. We'll show you more information." Additionally, if the current stock level is low, the product price will automatically increase.

[0336] Example prompt for generative AI model:

[0337] "How can I forecast sales trends and demand and create a dynamic interface that responds to user emotions?"

[0338] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0339] Step 1:

[0340] The server collects historical data for demand forecasting, including sales data from an SQL database, consumer behavior data, seasonality data, and market trend data through a web API. The input is data provided by various data sources, and the output is the collected raw data set.

[0341] Step 2:

[0342] The server preprocesses the collected data, which includes imputing missing data, correcting errors and outliers, and converting numerical data to a standard scale (ranging from 0 to 1). It also integrates data from different data sources. The input is the raw dataset collected in step 1, and the output is the preprocessed dataset.

[0343] Step 3:

[0344] The server trains a machine learning model based on the preprocessed data. First, it splits the data into training data and test data. Then it trains the model using a selected machine learning algorithm (e.g., LSTM, Random Forest, etc.). The input is the preprocessed dataset, and the output is the trained machine learning model.

[0345] Step 4:

[0346] The server uses the trained model to predict future demand. Test data is input to the trained model to predict future sales trends and demand. The inputs are the test data and the trained model, and the output is the predicted result data.

[0347] Step 5:

[0348] The server provides the prediction results to the user. The prediction results are visualized as line graphs or bar graphs and displayed on the user's device. The input is the prediction result data, and the output is the visual data displayed on the user's device.

[0349] Step 6:

[0350] The server uses a camera and microphone to recognize the user's emotions. It uses the EmotionEngine to analyze the user's facial expressions and voice to identify the user's emotions. The input is image and audio data obtained from the camera and microphone, and the output is the recognized emotion data.

[0351] Step 7:

[0352] The server dynamically changes the interface presentation based on the recognized emotion, for example, displaying additional information if the user is surprised, or a detailed explanation if the user is confused. The input is the recognized emotion data and the predicted result data, and the output is the adjusted interface.

[0353] Step 8:

[0354] The server dynamically sets prices based on the demand forecast results. When the forecast demand is high, the price is raised, and when there is excess inventory, the price is lowered. The input is the forecast result data and current inventory data, and the output is optimized price information.

[0355] Step 9:

[0356] The server sends new price information to the user's terminal in real time. The input is the optimized price information, and the output is the new price information displayed on the user's terminal.

[0357] 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.

[0358] 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.

[0359] 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.

[0360] [Second embodiment]

[0361] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0362] 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.

[0363] 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).

[0364] 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.

[0365] 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.

[0366] 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).

[0367] 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.

[0368] 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.

[0369] 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.

[0370] 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.

[0371] 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.

[0372] 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."

[0373] This invention provides a system that uses AI and machine learning to predict future sales trends and demand. The specific processing of the program used in this system is shown below.

[0374] 1. Data Collection

[0375] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and other related systems via APIs.

[0376] Example: Last year's monthly sales data is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API.

[0377] 2. Data Preprocessing

[0378] The server imputes missing data from the collected raw data, for example, by imputing the average value or by connecting previous and next values.

[0379] The server detects erroneous or outlier values ​​and corrects or removes them as necessary.

[0380] The server converts the numeric data to a standard scale (e.g., a range from 0 to 1) to allow comparison between different data.

[0381] The server combines and formats the data collected from different sources into a single integrated data set.

[0382] 3. Model Training

[0383] The server splits the dataset into training data and test data, typically 80% as training data and 20% as test data.

[0384] The server selects and initializes a machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) suitable for demand forecasting.

[0385] The server uses the selected machine learning algorithm to build a model using the training data, for example, inputting the data into an LSTM network to train a demand forecasting model.

[0386] 4. Demand forecasting

[0387] The server inputs test data using the trained machine learning model to predict future sales trends and demand.

[0388] Example: Use test data to forecast sales for the next month and check the results.

[0389] 5. Outputting the results

[0390] The server formats the prediction results appropriately and sends them to the user's device.

[0391] Example: Sales forecast results are visualized as line graphs and bar graphs and displayed on the user's dashboard.

[0392] 6. Dynamic Pricing

[0393] A server runs an algorithm that dynamically sets prices for products based on demand forecasts.

[0394] For example, increase prices by 10% before a period when demand is predicted to increase, or lower prices during periods of excess inventory to encourage sales.

[0395] The server takes into account the current stock amount and performs processing to lower the price if there is excess stock and to raise the price if there is low stock.

[0396] The server reflects the optimized pricing in the product database and transmits the new price information to the user's terminal.

[0397] The system's forecasting and pricing capabilities enable users to highly optimize inventory management, production scheduling, and dynamic pricing, thereby enabling companies to make more effective decisions in sales planning and maximize profits.

[0398] The processing flow will be explained below.

[0399] Step 1:

[0400] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and other related systems via APIs. For example, monthly sales data from the last year is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API.

[0401] Step 2:

[0402] The server imputes missing data from the collected data, for example by imputing the average value or by using values ​​before and after the data, and then detects erroneous or outlier values ​​and corrects or deletes them as necessary.

[0403] Step 3:

[0404] The server converts the numerical data to a standard scale (e.g., a range from 0 to 1) to enable comparisons between different data sets, and combines and formats data collected from different sources into a single integrated dataset.

[0405] Step 4:

[0406] The server splits the formatted dataset into training data and test data, typically using 80% of the data as training data and 20% as test data.

[0407] Step 5:

[0408] The server selects and initializes the machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) that is best suited for demand forecasting. For example, it selects the LSTM model and sets the necessary hyperparameters.

[0409] Step 6:

[0410] The server trains the model using the selected machine learning algorithm by providing training data. Specifically, it inputs past sales data into an LSTM network and builds a model to predict future sales.

[0411] Step 7:

[0412] The server inputs test data using the trained model to predict future sales trends and demand, for example, forecasting sales for the next month, and outputs the results as data.

[0413] Step 8:

[0414] The server then converts the forecast results into an appropriate format and sends them to the user's device. For example, the sales forecast results are visualized as line graphs or bar charts and displayed on the user's dashboard.

[0415] Step 9:

[0416] The server runs an algorithm that dynamically sets product prices based on predicted demand, for example, raising prices by 10% ahead of periods of predicted increased demand and lowering prices during periods of overstock.

[0417] Step 10:

[0418] The server takes into account the current inventory level and optimizes pricing, for example lowering the price if inventory is excessive and raising the price if inventory is low.

[0419] Step 11:

[0420] The server reflects the optimized pricing in the product database and transmits the new price information to the user's terminal in real time.

[0421] Through this series of steps, the system of the present invention can achieve highly accurate demand forecasting and dynamic pricing, optimizing a company's inventory management and sales planning.

[0422] Example 1

[0423] 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."

[0424] In recent years, retailers have found it difficult to accurately forecast sales trends and demand, which has led to problems in inventory management and pricing. In particular, the processes of data collection, data preprocessing, model training, and dynamic pricing required to predict demand fluctuations are complex, and a system for managing these processes efficiently and effectively is needed.

[0425] 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.

[0426] In this invention, the server includes: means for collecting past data from retailers and related systems; means for preprocessing the collected data by imputing missing data, detecting and removing outliers, standard scale conversion, and data combination; means for dividing the preprocessed data into training data and test data and selecting and training a machine learning algorithm suitable for demand forecasting; means for predicting future sales trends and demand using the trained machine learning model; means for visually displaying the prediction results and sending them to a user's terminal; and means for dynamically setting product prices based on the demand forecast results and inventory levels and updating the product database. This enables accurate prediction of demand fluctuations and efficient and effective inventory management and pricing.

[0427] "Data collection" is the process of obtaining historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and related systems.

[0428] "Data preprocessing" is the process of filling in missing data from collected raw data, detecting and removing outliers, and converting the data into a standard scale to put the data into a unified format.

[0429] "Machine learning algorithms" are mathematical models or methods used to predict future trends or demand based on a dataset. This invention uses LSTM, Random Forest, and XGBoost as examples.

[0430] "Training data" refers to the large dataset used to build and train machine learning models.

[0431] "Test data" refers to a dataset used to evaluate the performance of a trained machine learning model.

[0432] "Demand forecasting" is the process of using trained machine learning models to predict future sales trends and product demand.

[0433] "Dynamic pricing" is the process of automatically adjusting product prices based on demand forecasts and inventory levels.

[0434] "Product database" means a digital database for managing and storing product and pricing information.

[0435] "Visual display" is the process of presenting the results of prediction to a user in a visually easy-to-understand format such as a graph or table.

[0436] "User terminal" refers to a device, such as a PC or smartphone, that receives and displays prediction results and price information.

[0437] This invention provides a system that uses AI and machine learning to predict future sales trends and demand. This system operates in cooperation with a server, terminals, and users.

[0438] Data collection

[0439] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and related systems via APIs, specifically retrieving sales data from an SQL database and using the Google Analytics API to pull web browsing history data.

[0440] Example: Retrieve last year's monthly sales data using an SQL query and collect user web browsing history using the Google Analytics API.

[0441] Data Preprocessing

[0442] The server performs preprocessing on the collected raw data, including missing data imputation, outlier detection and removal, standard scale conversion, and data merging. Missing data is imputed using mean imputation, and outliers are filtered based on standard deviation. Data is standardized to a range of 0 to 1, and different data sources are combined into an integrated dataset.

[0443] Example: Use the Pandas library to impute NaN values ​​in a DataFrame, remove outliers, and perform normalization.

[0444] Model learning

[0445] The server splits the preprocessed data into training data and test data, selects and initializes a machine learning algorithm such as LSTM, Random Forest, or XGBoost, and builds and trains a model based on the training data.

[0446] Example: Using Sci-kit Learn, split the dataset into 80% training data and 20% test data, and build and train an LSTM model.

[0447] Demand forecasting

[0448] The server uses the trained model to predict future sales trends and demand based on the test data, and the prediction results are saved in a format such as a CSV file.

[0449] Example: Use test data to forecast sales for the next month and save the results to a file.

[0450] Output of results

[0451] The server visually displays the prediction results and sends them to the user's device. For example, it uses Matplotlib or Seaborn to draw line graphs and bar charts and display them on a dashboard.

[0452] Example: Generate a graph based on the prediction results and display it on a dashboard.

[0453] Dynamic Pricing

[0454] The server dynamically sets product prices based on demand forecasts and inventory levels, and updates the product database with this information. Based on a pricing algorithm, prices are increased during periods of increased demand and decreased when inventory levels are excessive.

[0455] Example: Raise prices by 10% when demand is predicted to increase, or lower prices when there is excess inventory.

[0456] Example prompt: "Use AI and machine learning to generate a Python script that uses the past year's sales data to forecast next month's sales and dynamically set pricing. Please include any libraries needed and specific code examples."

[0457] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0458] Step 1: Data collection

[0459] The server collects historical sales data, consumer behavior data, seasonal data, and market trend data from retailers and related systems. It queries an SQL database and retrieves web browsing history data using the Google Analytics API. The input is sales data, web click data, etc., and the output is a consolidated dataset containing these data. Specifically, the server securely collects data using API keys and authentication tokens.

[0460] Step 2: Data Preprocessing

[0461] The server preprocesses the collected data, including imputing missing data, detecting and removing outliers, converting to a standard scale, and combining different data sources. The input is raw data, and the output is preprocessed, clean data. The server uses the Pandas library to process the missing data, such as "df['Sales'].fillna(df['Sales'].mean(), inplace=True)", to filter outliers and convert to a standard scale. It also combines data from different data sources to generate a unified dataset.

[0462] Step 3: Split the data

[0463] The server splits the preprocessed data into training data and test data. The input is the clean dataset, and the output is the training data and test data. The server uses Sci-kit Learn and executes code such as "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)".

[0464] Step 4: Model training

[0465] The server uses the training data to train a machine learning model. The input is the training data, and the output is the trained model. The server selects an algorithm such as LSTM, Random Forest, or XGBoost, initializes it with code such as "from keras.models import Sequential" and "model = Sequential()", and trains the model with code such as "model.fit(X_train, y_train, epochs=50, batch_size=32, validation_split=0.2)".

[0466] Step 5: Demand forecast

[0467] The server uses the trained model to predict future sales trends and demand based on the test data. The input is the test data and the output is the prediction result. The server makes the prediction using "y_pred = model.predict(X_test)" and saves the prediction result in a format such as a CSV file.

[0468] Step 6: Output the results

[0469] The server visually displays the prediction results and sends them to the user's device. The input is the prediction results, and the output is the visually displayed data. The server uses Matplotlib and Seaborn to draw graphs, visualizes them with code such as "plt.plot(prediction result)", and displays them on a dashboard.

[0470] Step 7: Dynamic Pricing

[0471] The server dynamically sets product prices based on demand forecast results and inventory levels, and reflects this information in the product database. The inputs are forecast results and inventory data, and the output is dynamically set price information. The server adjusts prices during periods of increased demand, such as "item_price = 1.10", and updates the new price information in the product database. Conditional branching such as "if inventory level > threshold: price = 0.90" and "elif inventory level < threshold: price = 1.10" is used. Specifically, the server updates price information via an SQL query and sends the new price information to the user's device.

[0472] (Application example 1)

[0473] 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."

[0474] Conventional demand forecasting systems require a lot of time and resources to compile data and make forecasts, making it difficult to respond quickly to fluctuations in demand. Furthermore, there are limited ways to present forecast results in a visually understandable manner, making it difficult for managers and store managers to make appropriate decisions. Furthermore, when it comes to dynamic pricing, there is a lack of functionality to automatically present effective strategies based on demand forecasts.

[0475] 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.

[0476] In this invention, the server includes means for collecting past data for demand forecasting, means for preprocessing the collected data, means for training a machine learning model based on the preprocessed data, means for forecasting future demand using the trained machine learning model, means for providing the forecast results to a user, means for dynamically setting prices based on the demand forecast results, means for predicting sales trends and proposing appropriate price increases or decreases, and means for visualizing the forecast results, thereby enabling store managers to quickly and effectively perform demand forecasting and optimize dynamic pricing and inventory management.

[0477] "Demand forecasting" is the prediction of future customer demand and sales volume based on past data.

[0478] "Data collection" refers to the collection of historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and related systems.

[0479] "Data preprocessing" refers to processing collected raw data, such as filling in missing data, correcting or deleting error values ​​and outliers, converting numerical data to a standard scale, and integrating different data.

[0480] A "machine learning model" is an algorithm that is trained to perform tasks such as prediction or classification using collected and preprocessed data as input.

[0481] "Dynamic pricing" refers to adjusting the price of goods and services in real time in response to changes in supply and demand.

[0482] "Collected Data" refers to historical sales data, consumer behavior data, seasonal data, and market trend data obtained from retailers and related systems via APIs or other means.

[0483] A "machine learning algorithm" is a method or procedure for performing machine learning tasks such as demand forecasting and classification based on various data, for example, LSTM, Random Forest, or XGBoost.

[0484] "Sales trend forecasting" refers to predicting future sales figures and sales trends based on past sales data and other related information.

[0485] A "price increase or decrease proposal" is a specific recommendation to increase or decrease the price of a product based on projected demand and inventory availability.

[0486] "Visualization of forecast results" refers to displaying predicted demand and sales trends in a visual format such as graphs and charts.

[0487] This invention relates to a demand forecasting and dynamic pricing system that utilizes AI and machine learning to predict future demand based on a wide range of data, including past sales data, and dynamically sets prices for products and services based on the results.

[0488] First, the server collects past sales data, consumer behavior data, seasonal data, and market trend data from retailers and related systems via API. For example, monthly sales data from the last fiscal year is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API. The data collected in this way is stored on the server as raw data.

[0489] Next, the server preprocesses the collected data. Missing data is imputed with the average value, and errors and outliers are detected and corrected or removed. Collected data from different data sources is integrated, converted to a standard scale (e.g., a range from 0 to 1), and organized into a unified dataset. The main software used in this process includes Python's Pandas and Numpy.

[0490] The server trains a machine learning model based on the preprocessed data. The dataset is divided into training data and test data, and a machine learning algorithm such as random forest or LSTM is selected to build the model. For example, past sales data and consumer behavior data are input to train the LSTM network. Libraries such as Scikit-learn and TensorFlow are used in this process.

[0491] The server uses the trained machine learning model to predict future demand. Test data can be input into the model to predict future sales trends. The prediction results are sent to the user's device and displayed in a visualized format, such as a line graph or bar chart. The results are presented in an intuitive and easy-to-understand format for managers and store administrators.

[0492] The server also has a means for dynamically setting prices based on the results of demand forecasts. It automatically adjusts prices by raising them during periods of increased demand and lowering them when there is excess inventory. Prices are dynamically set in conjunction with current inventory levels, promoting sales. This algorithm is reflected in the product database, and updated price information is sent to users' devices in real time.

[0493] For example, when weekly sales data, consumer behavior data, and market trend data are input, the system will suggest appropriate price increases or decreases based on the predicted sales trends for the following month.An example of a prompt to be input into the generative AI model is, "Based on last week's sales data, consumer behavior data, and market trend data, please suggest predicted sales trends and dynamic pricing for the next two weeks."

[0494] In this way, the present invention provides an efficient tool for predicting future sales trends and dynamic pricing in brick-and-mortar stores.

[0495] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0496] Step 1: Data collection

[0497] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and related systems via APIs. This may include, for example, retrieving monthly sales data from an SQL database or collecting customer web browsing history through the Google Analytics API. The input of this step is the data source, and the output is the collected raw data.

[0498] Step 2: Data Preprocessing

[0499] The server preprocesses the collected raw data by first imputing missing data with the mean value, then detecting and correcting or removing erroneous values ​​and outliers, normalizing numerical data to a range of 0 to 1, and integrating data collected from different data sources. The input of this step is the collected raw data, and the output is a preprocessed dataset.

[0500] Step 3: Train the model

[0501] The server uses the preprocessed data to train a machine learning model. First, it splits the dataset into training data and test data. Then it selects an algorithm, such as random forest or LSTM, and builds a model using the training data. The input of this step is the preprocessed data, and the output is a trained machine learning model.

[0502] Step 4: Demand forecast

[0503] The server uses the trained machine learning model to predict future demand. Specifically, it inputs test data into the model and calculates future sales trends. The inputs for this step are the test data and the trained model, and the output is predicted demand data.

[0504] Step 5: Providing prediction results

[0505] The server provides the forecast results to the user. For example, it displays the forecasted demand data on the user's terminal in a visual format such as a line graph or bar graph. The input of this step is the forecasted demand data, and the output is the visualized forecast results.

[0506] Step 6: Dynamic Pricing

[0507] The server dynamically sets prices based on the demand forecast results. Specifically, it raises product prices before the period when demand is predicted to increase, and lowers prices when there is excess inventory. It also takes into account current inventory levels and raises prices when inventory is low. The inputs to this step are predicted demand data and inventory data, and the output is dynamically set price information.

[0508] Step 7: Sales trend forecast and proposals

[0509] The server proposes price increases or decreases based on future sales trends. For example, by inputting weekly sales data, consumer behavior data, and market trend data, the server proposes appropriate price adjustments based on the predicted sales trends for the next month. The inputs for this step are the collected data and predicted sales trends, and the output is a specific price adjustment proposal.

[0510] Step 8: Visualize the prediction results

[0511] The server visualizes the predicted demand and sales trends. It displays the forecast results in line graphs and bar charts, and communicates them in an intuitive and easy-to-understand format to the user. The input to this step is the predicted demand data and sales trend data, and the output is visualized graphs and charts.

[0512] 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.

[0513] This invention combines a system that uses AI and machine learning to predict future sales trends and demand with an emotion engine that recognizes user emotions. The specific processing of the program used in this system is shown below.

[0514] 1. Data Collection

[0515] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and other related systems via APIs. For example, monthly sales data from the last year is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API.

[0516] 2. Data Preprocessing

[0517] The server imputes missing data from the collected data, for example by imputing the average value or by using values ​​before and after the data, and then detects erroneous or outlier values ​​and corrects or deletes them as necessary.

[0518] The server converts the numerical data to a standard scale (e.g., a range from 0 to 1) to enable comparisons between different data sets, and combines and formats data collected from different sources into a single integrated dataset.

[0519] 3. Model Training

[0520] The server splits the formatted dataset into training data and test data, typically using 80% of the data as training data and 20% as test data.

[0521] The server selects and initializes the machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) that is best suited for demand forecasting. For example, it selects the LSTM model and sets the necessary hyperparameters.

[0522] The server trains the model using the selected machine learning algorithm by providing training data. Specifically, it inputs past sales data into an LSTM network and builds a model to predict future sales.

[0523] 4. Demand forecasting

[0524] The server inputs test data using the trained model to predict future sales trends and demand, for example, forecasting sales for the next month, and outputs the results as data.

[0525] 5. Outputting the results

[0526] The server then converts the forecast results into an appropriate format and sends them to the user's device. For example, the sales forecast results are visualized as line graphs or bar charts and displayed on the user's dashboard.

[0527] The emotion engine uses a camera and microphone to analyze the user's facial expressions and voice when presenting results, and dynamically changes the way prediction results are presented based on the recognized emotions.

[0528] For example: If the user is surprised, show them some additional information to help them calm down. If the user is confused, show them a detailed explanation.

[0529] 6. Dynamic Pricing

[0530] The server runs an algorithm that dynamically sets product prices based on predicted demand, for example, raising prices by 10% ahead of periods of predicted increased demand and lowering prices during periods of overstock.

[0531] The server takes into account the current inventory level and optimizes pricing, for example lowering the price if inventory is excessive and raising the price if inventory is low.

[0532] The server reflects the optimized pricing in the product database and transmits the new price information to the user's terminal in real time.

[0533] In addition to the demand forecasting and dynamic pricing functions provided by this system, by integrating user emotion recognition with an emotion engine, it is possible to provide a more intuitive and adaptive interface for users, which will enable companies to improve not only inventory management and production scheduling but also the user experience.

[0534] The processing flow will be explained below.

[0535] Step 1:

[0536] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and other related systems via APIs. For example, monthly sales data from the last year is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API.

[0537] Step 2:

[0538] The server imputes missing data from the collected data, for example by imputing the average value or by using values ​​before and after the data, and then detects erroneous or outlier values ​​and corrects or deletes them as necessary.

[0539] Step 3:

[0540] The server converts numerical data to a standard scale (e.g., a range from 0 to 1) to allow comparisons between different data sets, and combines and formats data collected from different sources into a single integrated dataset.

[0541] Step 4:

[0542] The server splits the formatted dataset into training data and test data, typically using 80% of the data as training data and 20% as test data.

[0543] Step 5:

[0544] The server selects and initializes the machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) that is best suited for demand forecasting. For example, it selects the LSTM model and sets the necessary hyperparameters.

[0545] Step 6:

[0546] The server trains the model using the selected machine learning algorithm by providing training data. Specifically, it inputs past sales data into an LSTM network and builds a model to predict future sales.

[0547] Step 7:

[0548] The server inputs test data using the trained model to predict future sales trends and demand, for example, forecasting sales for the next month, and outputs the results as data.

[0549] Step 8:

[0550] The server then converts the forecast results into an appropriate format and sends them to the user's device. For example, the sales forecast results are visualized as line graphs or bar charts and displayed on the user's dashboard.

[0551] Step 9:

[0552] The device uses a camera and microphone to collect the user's facial expressions and voice data, which the emotion engine analyzes to recognize the user's emotions, such as whether the user is confused or surprised.

[0553] Step 10:

[0554] The server dynamically changes how the predictions are presented based on the recognized emotion: for example, if the user is surprised, it may display additional information to calm them down, or if the user is confused, it may display additional detailed explanations or instructions.

[0555] Step 11:

[0556] The server runs an algorithm that dynamically sets product prices based on predicted demand, for example, raising prices by 10% ahead of periods of predicted increased demand and lowering prices during periods of overstock.

[0557] Step 12:

[0558] The server takes into account the current inventory level and optimizes pricing, for example lowering the price if inventory is excessive and raising the price if inventory is low.

[0559] Step 13:

[0560] The server reflects the optimized pricing in the product database and transmits the new price information to the user's terminal in real time.

[0561] Step 14:

[0562] Users can check new price information and forecast results through their devices, and can adjust their inventory management and sales strategies accordingly.

[0563] Through this series of steps, the system of the present invention achieves highly accurate demand forecasting and dynamic pricing, and also provides an adaptive interface that responds to user emotions, allowing companies to optimize all aspects of inventory management, product production, and sales strategies.

[0564] Example 2

[0565] 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."

[0566] Conventional demand forecasting systems use past data to predict future demand, but they are unable to provide an interface that takes user emotions into account, resulting in a poor user experience. In particular, because the user's psychological state is not reflected when presenting the forecast results, there is a lack of support for the user to understand the results and take action. Conventional systems also lack the ability to dynamically set prices.

[0567] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0568] In this invention, the server includes means for collecting past data for demand forecasting, means for preprocessing the collected data, means for training a machine learning model based on the preprocessed data, means for forecasting future demand using the trained machine learning model, means for providing the forecast results to a user, means for recognizing the user's emotions, means for dynamically presenting the forecast results based on the user's emotions, and means for dynamically setting prices based on the demand forecast results. This makes it possible to provide an interface that reflects the user's emotions, improving the user experience and providing more support for understanding and taking action on the demand forecast results. Furthermore, dynamic pricing enables flexible price management in response to market fluctuations.

[0569] "Demand forecasting" is the process of predicting future consumer demand based on past and current data.

[0570] "Data Collection Measures" refers to the ability to collect necessary data from retailers and other related systems.

[0571] "Data preprocessing means" refers to functions that fill in missing values ​​in collected data, correct or delete error values ​​or outliers, and perform scaling and integration.

[0572] "Machine learning model training means" refers to a function that uses preprocessed data to train a machine learning algorithm and create a predictive model.

[0573] "Demand forecasting means" refers to the ability to forecast future demand using a trained machine learning model.

[0574] "Means of providing to users" refers to the function of providing prediction results to users in an appropriate format and presenting them in an easy-to-understand manner.

[0575] "Emotion recognition means" refers to a function that analyzes the user's facial expressions and voice to recognize the user's emotional state.

[0576] "Dynamic presentation means" refers to a function that dynamically changes the way prediction results are presented based on the recognized user emotions.

[0577] "Dynamic pricing means" refers to a function that dynamically sets product prices in real time based on demand forecast results.

[0578] This invention combines a system that uses AI and machine learning to predict future sales trends and demand with an emotion engine that recognizes user emotions. The specific processing of the program used in this system is described in detail below.

[0579] First, the server collects data. It collects past sales data, consumer behavior data, seasonal data, market trend data, etc. from retailers and related systems. Specifically, it retrieves last year's monthly sales data from an SQL database and collects users' web browsing history using the Google Analytics API. The data collected in this way includes information from a variety of sources.

[0580] Next, the server performs data preprocessing. Missing data is imputed from the collected data, and errors and outliers are detected and corrected or removed. Specific processing methods include average imputation and imputation based on surrounding values. All numerical data is also converted to a standard scale to enable comparisons between different data sets. Data collected from different sources is combined into a single integrated dataset and formatted.

[0581] The server then performs model training. The formatted dataset is divided into training data and test data, and the machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) optimal for demand forecasting is selected and initialized. For example, an LSTM model is selected and the necessary hyperparameters are set. Training data is input using this selected algorithm to train the model. Specifically, past sales data is input into the LSTM network to build a model for predicting future sales.

[0582] Next, the server performs demand forecasting. It uses the trained model to input test data and predict future sales trends and demand. For example, it predicts sales for the next month and outputs the results as data.

[0583] The prediction results are provided from the server to the user's device in an appropriate format. The prediction results are visualized in line graphs and bar charts and displayed on the user's dashboard. The emotion engine uses a camera and microphone to analyze the user's facial expressions and voice when presenting the results, recognizing the user's emotions. Based on the recognized emotions, the way the prediction results are presented is dynamically changed. For example, if the user is surprised, additional information to calm them down is displayed, and if the user is confused, a detailed explanation is displayed.

[0584] Dynamic pricing is also an important feature of this system. The server runs an algorithm that dynamically sets product prices based on predicted demand. For example, the price can be raised by 10% before a period when demand is predicted to increase, and lowered when there is excess inventory. Furthermore, pricing is optimized taking into account current inventory levels. For example, if there is excess inventory, the price is lowered, and if inventory is low, the price is raised. The optimized pricing is reflected in the product database, and new price information is sent to users' devices in real time.

[0585] The system integrates demand forecasting and dynamic pricing functions, as well as user emotion recognition through an emotion engine, to provide a more intuitive and adaptive interface for users, helping companies not only to optimize inventory management and production schedules, but also to improve the user experience.

[0586] Specific examples

[0587] For example, the server retrieves last year's sales data from an SQL database and collects users' web browsing history using the Google Analytics API. It then uses mean imputation to impute missing data and scales the data to a range of 0 to 1 using MinMaxScaler. It selects an LSTM model and trains it using the training data with the fit() function. It uses the predict() function to predict next month's sales and displays the results in a line graph. The emotion engine analyzes emotions from the user's facial expressions and voice and dynamically changes the way the prediction results are presented. Based on demand, the server may raise prices by 10% or optimize prices, update the product database, and notify the user's device of the new price information.

[0588] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0589] Program processing flow and specific explanation

[0590] Step 1: Data collection

[0591] Input: Database connection information for retailers and related systems, API keys

[0592] Specific behavior:

[0593] The server uses an SQL query to retrieve last year's monthly sales data from the database.

[0594] The server uses the Google Analytics API to collect the user's web browsing history for the past six months.

[0595] Output: Collected dataset of historical sales data and consumer behavior data

[0596] Step 2: Data Preprocessing

[0597] Input: raw data collected, missing data, data containing error values ​​and outliers

[0598] Specific behavior:

[0599] The server generates a data frame using the Pandas library and performs imputation of missing values ​​(e.g., mean imputation).

[0600] The server runs an outlier detection algorithm to detect and remove or correct outliers.

[0601] The server applies MinMaxScaler to scale the numeric data to the range 0 to 1.

[0602] The server combines and formats the data collected from different sources into a single integrated data set.

[0603] Output: Preprocessed consolidated dataset

[0604] Step 3: Model training

[0605] Input: Preprocessed unified dataset, dataset split into training and test data

[0606] Specific behavior:

[0607] The server splits the combined dataset into training data (80%) and test data (20%) using the Scikit-learn or Keras libraries.

[0608] The server selects an LSTM model and sets hyperparameters (e.g., learning rate, number of epochs).

[0609] The server uses the training data to train the LSTM model using the fit() function.

[0610] Output: A trained LSTM model

[0611] Step 4: Demand forecast

[0612] Input: trained model, test data

[0613] Specific behavior:

[0614] The server uses the predict() function to forecast future sales trends and demand based on the test data.

[0615] Specifically, it generates sales forecast results for the next month.

[0616] Output: Future sales forecast results

[0617] Step 5: Output the results

[0618] Input: Future sales forecast results

[0619] Specific behavior:

[0620] The server uses the Matplotlib and Seaborn libraries to visualize the forecast results in line and bar graphs.

[0621] The server sends the visualized data to the user's device and displays it on a dashboard.

[0622] Output: Visualized data for dashboard display

[0623] Step 6: Emotion Recognition

[0624] Input: User's facial expressions, voice data

[0625] Specific behavior:

[0626] The emotion engine analyzes facial and voice data collected from the camera and microphone to recognize the user's emotional state.

[0627] The emotion engine uses facial recognition and voice analysis algorithms.

[0628] Output: Recognized emotional state of the user

[0629] Step 7: Dynamic Presentation

[0630] Input: Recognized user emotional state, future sales forecast results

[0631] Specific behavior:

[0632] The server dynamically changes the way predictions are presented based on the user's perceived emotions: if the user is surprised, it displays additional reassuring information, and if the user is confused, it displays a detailed explanation.

[0633] Output: Dynamically changed presentation results based on user emotion

[0634] Step 8: Dynamic Pricing

[0635] Input: Future sales forecast results, current inventory amount

[0636] Specific behavior:

[0637] The server dynamically sets product prices based on predicted demand, for example by raising prices by 10% before periods when demand is predicted to increase, and lowering prices when there is excess inventory.

[0638] The server optimizes pricing by taking into account inventory levels. Specifically, if inventory is excessive, the price is reduced by 5%, and if inventory is low, the price is increased by 5%.

[0639] Output: Optimized pricing information

[0640] Step 9: Update and submit pricing information

[0641] Input: Optimized Pricing Information

[0642] Specific behavior:

[0643] The server updates the product database and transmits new price information to the user terminal in real time.

[0644] Output: Updated price information, real-time notification to user terminal

[0645] (Application example 2)

[0646] 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."

[0647] Conventional demand forecasting systems primarily use historical data and machine learning to predict future sales trends, but they are unable to consider the user's emotional state. As a result, it is difficult to provide optimal interfaces and adaptive feedback for users. In particular, the lack of emotion recognition makes it difficult for users to accept and understand the prediction results, resulting in reduced usability. Furthermore, while dynamic pricing is automatically set based on demand forecasts, it fails to sufficiently increase user satisfaction because pricing based on user emotions and behavior is not taken into account.

[0648] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0649] In this invention, the server includes means for collecting past data for demand forecasting, means for preprocessing the collected data, means for training a machine learning model based on the preprocessed data, means for forecasting future demand using the trained machine learning model, means for providing the forecast results to a user, means for dynamically setting prices based on the demand forecast results, and means for recognizing the user's emotions and dynamically changing the presentation method of the forecast results and the interface based on the emotions. This makes it possible to provide more intuitive and adaptive feedback while taking the user's emotions into consideration. Dynamic pricing based on the user's emotions is also realized, thereby improving user satisfaction.

[0650] "Demand forecasting" refers to analyzing past data and predicting future sales trends and demand.

[0651] "Means for collection" refers to a function that allows necessary data to be acquired from various data sources.

[0652] "Means for data preprocessing" refers to the function of completing missing data and correcting outliers in collected data, and processing it to prepare it in a format that can be analyzed.

[0653] A "machine learning model" is a trained algorithm used to predict future demand based on historical data.

[0654] "Training" refers to the process of inputting data into a machine learning model and adjusting the model's parameters to improve its predictive accuracy.

[0655] A "demand forecasting means" is one that has the ability to forecast future sales trends and demand using trained machine learning models.

[0656] "Means for providing to the user" refers to a function that provides an interface for displaying prediction results to the user.

[0657] "Dynamic pricing" refers to the ability to adjust product prices in real time based on predicted demand.

[0658] "Means for recognizing user emotions" refers to technology that uses a camera and microphone to analyze the user's facial expressions and voice to identify their emotional state.

[0659] "Means for dynamically changing the interface presentation method" refers to a function that adaptively changes the screen display and information presentation method based on the user's emotions.

[0660] A system for implementing the present invention is configured as follows.

[0661] First, the server collects historical data for demand forecasting. The collected data includes past sales data, consumer behavior data, seasonal data, and market trend data. This data is collected by querying an SQL database or retrieving data through APIs.

[0662] The server then preprocesses the collected data, which includes imputing missing data, correcting errors and outliers, converting numerical data to a standard scale, and integrating data from different sources. This process prepares the data in a format that can be analyzed.

[0663] The server then trains a machine learning model based on the preprocessed data. The machine learning model to be used may be LSTM, Random Forest, XGBoost, or other models. The data is separated into training data and test data, and the model is trained using the selected algorithm. This process builds a highly accurate model for predicting future demand.

[0664] The server uses the trained model to predict future demand, and the predicted demand data is provided to the user. For example, the sales forecast results are displayed on the user's device as a line graph or bar chart.

[0665] The server also uses a camera and microphone to recognize the user's emotions. Using audio and image analysis software such as EmotionEngine, the server analyzes the user's facial expressions and voice to identify their emotions. Based on this identified emotion, the server dynamically changes the way prediction results are presented and the interface. For example, if the user is surprised, the server displays additional information along with the prediction result.

[0666] The server also dynamically sets prices based on the results of demand forecasts. Prices are raised if demand is predicted to increase, and lowered if there is excess inventory. This allows companies to optimize inventory management and pricing.

[0667] This system provides the following advantages: Improved usability, as users can intuitively understand the information they want; and Dynamic pricing, which maximizes profits for businesses. This is a specific embodiment of the present invention.

[0668] For example, if the user is surprised, a message will appear on the screen saying, "Don't worry. We'll show you more information." Additionally, if the current stock level is low, the product price will automatically increase.

[0669] Example prompt for generative AI model:

[0670] "How can I forecast sales trends and demand and create a dynamic interface that responds to user emotions?"

[0671] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0672] Step 1:

[0673] The server collects historical data for demand forecasting, including sales data from an SQL database, consumer behavior data, seasonality data, and market trend data through a web API. The input is data provided by various data sources, and the output is the collected raw data set.

[0674] Step 2:

[0675] The server preprocesses the collected data, which includes imputing missing data, correcting errors and outliers, and converting numerical data to a standard scale (ranging from 0 to 1). It also integrates data from different data sources. The input is the raw dataset collected in step 1, and the output is the preprocessed dataset.

[0676] Step 3:

[0677] The server trains a machine learning model based on the preprocessed data. First, it splits the data into training data and test data. Then it trains the model using a selected machine learning algorithm (e.g., LSTM, Random Forest, etc.). The input is the preprocessed dataset, and the output is the trained machine learning model.

[0678] Step 4:

[0679] The server uses the trained model to predict future demand. Test data is input to the trained model to predict future sales trends and demand. The inputs are the test data and the trained model, and the output is the predicted result data.

[0680] Step 5:

[0681] The server provides the prediction results to the user. The prediction results are visualized as line graphs or bar graphs and displayed on the user's device. The input is the prediction result data, and the output is the visual data displayed on the user's device.

[0682] Step 6:

[0683] The server uses a camera and microphone to recognize the user's emotions. It uses the EmotionEngine to analyze the user's facial expressions and voice to identify the user's emotions. The input is image and audio data obtained from the camera and microphone, and the output is the recognized emotion data.

[0684] Step 7:

[0685] The server dynamically changes the interface presentation based on the recognized emotion, for example, displaying additional information if the user is surprised, or a detailed explanation if the user is confused. The input is the recognized emotion data and the predicted result data, and the output is the adjusted interface.

[0686] Step 8:

[0687] The server dynamically sets prices based on the demand forecast results. When the forecast demand is high, the price is raised, and when there is excess inventory, the price is lowered. The input is the forecast result data and current inventory data, and the output is optimized price information.

[0688] Step 9:

[0689] The server sends new price information to the user's terminal in real time. The input is the optimized price information, and the output is the new price information displayed on the user's terminal.

[0690] 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.

[0691] 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.

[0692] 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.

[0693] [Third embodiment]

[0694] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0695] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[0696] 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).

[0697] 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.

[0698] 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.

[0699] 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).

[0700] 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.

[0701] 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.

[0702] 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.

[0703] 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.

[0704] 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.

[0705] 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."

[0706] This invention provides a system that uses AI and machine learning to predict future sales trends and demand. The specific processing of the program used in this system is shown below.

[0707] 1. Data Collection

[0708] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and other related systems via APIs.

[0709] Example: Last year's monthly sales data is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API.

[0710] 2. Data Preprocessing

[0711] The server imputes missing data from the collected raw data, for example, by imputing the average value or by connecting previous and next values.

[0712] The server detects erroneous or outlier values ​​and corrects or removes them as necessary.

[0713] The server converts the numeric data to a standard scale (e.g., a range from 0 to 1) to allow comparison between different data.

[0714] The server combines and formats the data collected from different sources into a single integrated data set.

[0715] 3. Model Training

[0716] The server splits the dataset into training data and test data, typically 80% as training data and 20% as test data.

[0717] The server selects and initializes a machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) suitable for demand forecasting.

[0718] The server uses the selected machine learning algorithm to build a model using the training data, for example, inputting the data into an LSTM network to train a demand forecasting model.

[0719] 4. Demand forecasting

[0720] The server inputs test data using the trained machine learning model to predict future sales trends and demand.

[0721] Example: Use test data to forecast sales for the next month and check the results.

[0722] 5. Outputting the results

[0723] The server formats the prediction results appropriately and sends them to the user's device.

[0724] Example: Sales forecast results are visualized as line graphs and bar graphs and displayed on the user's dashboard.

[0725] 6. Dynamic Pricing

[0726] A server runs an algorithm that dynamically sets prices for products based on demand forecasts.

[0727] For example, increase prices by 10% before a period when demand is predicted to increase, or lower prices during periods of excess inventory to encourage sales.

[0728] The server takes into account the current stock amount and performs processing to lower the price if there is excess stock and to raise the price if there is low stock.

[0729] The server reflects the optimized pricing in the product database and transmits the new price information to the user's terminal.

[0730] The system's forecasting and pricing capabilities enable users to highly optimize inventory management, production scheduling, and dynamic pricing, thereby enabling companies to make more effective decisions in sales planning and maximize profits.

[0731] The processing flow will be explained below.

[0732] Step 1:

[0733] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and other related systems via APIs. For example, monthly sales data from the last year is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API.

[0734] Step 2:

[0735] The server imputes missing data from the collected data, for example by imputing the average value or by using values ​​before and after the data, and then detects erroneous or outlier values ​​and corrects or deletes them as necessary.

[0736] Step 3:

[0737] The server converts the numerical data to a standard scale (e.g., a range from 0 to 1) to enable comparisons between different data sets, and combines and formats data collected from different sources into a single integrated dataset.

[0738] Step 4:

[0739] The server splits the formatted dataset into training data and test data, typically using 80% of the data as training data and 20% as test data.

[0740] Step 5:

[0741] The server selects and initializes the machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) that is best suited for demand forecasting. For example, it selects the LSTM model and sets the necessary hyperparameters.

[0742] Step 6:

[0743] The server trains the model using the selected machine learning algorithm by providing training data. Specifically, it inputs past sales data into an LSTM network and builds a model to predict future sales.

[0744] Step 7:

[0745] The server inputs test data using the trained model to predict future sales trends and demand, for example, forecasting sales for the next month, and outputs the results as data.

[0746] Step 8:

[0747] The server then converts the forecast results into an appropriate format and sends them to the user's device. For example, the sales forecast results are visualized as line graphs or bar charts and displayed on the user's dashboard.

[0748] Step 9:

[0749] The server runs an algorithm that dynamically sets product prices based on predicted demand, for example, raising prices by 10% ahead of periods of predicted increased demand and lowering prices during periods of overstock.

[0750] Step 10:

[0751] The server takes into account the current inventory level and optimizes pricing, for example lowering the price if inventory is excessive and raising the price if inventory is low.

[0752] Step 11:

[0753] The server reflects the optimized pricing in the product database and transmits the new price information to the user's terminal in real time.

[0754] Through this series of steps, the system of the present invention can achieve highly accurate demand forecasting and dynamic pricing, optimizing a company's inventory management and sales planning.

[0755] Example 1

[0756] 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."

[0757] In recent years, retailers have found it difficult to accurately forecast sales trends and demand, which has led to problems in inventory management and pricing. In particular, the processes of data collection, data preprocessing, model training, and dynamic pricing required to predict demand fluctuations are complex, and a system for managing these processes efficiently and effectively is needed.

[0758] 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.

[0759] In this invention, the server includes: means for collecting past data from retailers and related systems; means for preprocessing the collected data by imputing missing data, detecting and removing outliers, standard scale conversion, and data combination; means for dividing the preprocessed data into training data and test data and selecting and training a machine learning algorithm suitable for demand forecasting; means for predicting future sales trends and demand using the trained machine learning model; means for visually displaying the prediction results and sending them to a user's terminal; and means for dynamically setting product prices based on the demand forecast results and inventory levels and updating the product database. This enables accurate prediction of demand fluctuations and efficient and effective inventory management and pricing.

[0760] "Data collection" is the process of obtaining historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and related systems.

[0761] "Data preprocessing" is the process of filling in missing data from collected raw data, detecting and removing outliers, and converting the data into a standard scale to put the data into a unified format.

[0762] "Machine learning algorithms" are mathematical models or methods used to predict future trends or demand based on a dataset. This invention uses LSTM, Random Forest, and XGBoost as examples.

[0763] "Training data" refers to the large dataset used to build and train machine learning models.

[0764] "Test data" refers to a dataset used to evaluate the performance of a trained machine learning model.

[0765] "Demand forecasting" is the process of using trained machine learning models to predict future sales trends and product demand.

[0766] "Dynamic pricing" is the process of automatically adjusting product prices based on demand forecasts and inventory levels.

[0767] "Product database" means a digital database for managing and storing product and pricing information.

[0768] "Visual display" is the process of presenting the results of prediction to a user in a visually easy-to-understand format such as a graph or table.

[0769] "User terminal" refers to a device, such as a PC or smartphone, that receives and displays prediction results and price information.

[0770] This invention provides a system that uses AI and machine learning to predict future sales trends and demand. This system operates in cooperation with a server, terminals, and users.

[0771] Data collection

[0772] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and related systems via APIs, specifically retrieving sales data from an SQL database and using the Google Analytics API to pull web browsing history data.

[0773] Example: Retrieve last year's monthly sales data using an SQL query and collect user web browsing history using the Google Analytics API.

[0774] Data Preprocessing

[0775] The server performs preprocessing on the collected raw data, including missing data imputation, outlier detection and removal, standard scale conversion, and data merging. Missing data is imputed using mean imputation, and outliers are filtered based on standard deviation. Data is standardized to a range of 0 to 1, and different data sources are combined into an integrated dataset.

[0776] Example: Use the Pandas library to impute NaN values ​​in a DataFrame, remove outliers, and perform normalization.

[0777] Model learning

[0778] The server splits the preprocessed data into training data and test data, selects and initializes a machine learning algorithm such as LSTM, Random Forest, or XGBoost, and builds and trains a model based on the training data.

[0779] Example: Using Sci-kit Learn, split the dataset into 80% training data and 20% test data, and build and train an LSTM model.

[0780] Demand forecasting

[0781] The server uses the trained model to predict future sales trends and demand based on the test data, and the prediction results are saved in a format such as a CSV file.

[0782] Example: Use test data to forecast sales for the next month and save the results to a file.

[0783] Output of results

[0784] The server visually displays the prediction results and sends them to the user's device. For example, it uses Matplotlib or Seaborn to draw line graphs and bar charts and display them on a dashboard.

[0785] Example: Generate a graph based on the prediction results and display it on a dashboard.

[0786] Dynamic Pricing

[0787] The server dynamically sets product prices based on demand forecasts and inventory levels, and updates the product database with this information. Based on a pricing algorithm, prices are increased during periods of increased demand and decreased when inventory levels are excessive.

[0788] Example: Raise prices by 10% when demand is predicted to increase, or lower prices when there is excess inventory.

[0789] Example prompt: "Use AI and machine learning to generate a Python script that uses the past year's sales data to forecast next month's sales and dynamically set pricing. Please include any libraries needed and specific code examples."

[0790] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0791] Step 1: Data collection

[0792] The server collects historical sales data, consumer behavior data, seasonal data, and market trend data from retailers and related systems. It queries an SQL database and retrieves web browsing history data using the Google Analytics API. The input is sales data, web click data, etc., and the output is a consolidated dataset containing these data. Specifically, the server securely collects data using API keys and authentication tokens.

[0793] Step 2: Data Preprocessing

[0794] The server preprocesses the collected data, including imputing missing data, detecting and removing outliers, converting to a standard scale, and combining different data sources. The input is raw data, and the output is preprocessed, clean data. The server uses the Pandas library to process the missing data, such as "df['Sales'].fillna(df['Sales'].mean(), inplace=True)", to filter outliers and convert to a standard scale. It also combines data from different data sources to generate a unified dataset.

[0795] Step 3: Split the data

[0796] The server splits the preprocessed data into training data and test data. The input is the clean dataset, and the output is the training data and test data. The server uses Sci-kit Learn and executes code such as "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)".

[0797] Step 4: Model training

[0798] The server uses the training data to train a machine learning model. The input is the training data, and the output is the trained model. The server selects an algorithm such as LSTM, Random Forest, or XGBoost, initializes it with code such as "from keras.models import Sequential" and "model = Sequential()", and trains the model with code such as "model.fit(X_train, y_train, epochs=50, batch_size=32, validation_split=0.2)".

[0799] Step 5: Demand forecast

[0800] The server uses the trained model to predict future sales trends and demand based on the test data. The input is the test data and the output is the prediction result. The server makes the prediction using "y_pred = model.predict(X_test)" and saves the prediction result in a format such as a CSV file.

[0801] Step 6: Output the results

[0802] The server visually displays the prediction results and sends them to the user's device. The input is the prediction results, and the output is the visually displayed data. The server uses Matplotlib and Seaborn to draw graphs, visualizes them with code such as "plt.plot(prediction result)", and displays them on a dashboard.

[0803] Step 7: Dynamic Pricing

[0804] The server dynamically sets product prices based on demand forecast results and inventory levels, and reflects this information in the product database. The inputs are forecast results and inventory data, and the output is dynamically set price information. The server adjusts prices during periods of increased demand, such as "item_price = 1.10", and updates the new price information in the product database. Conditional branching such as "if inventory level > threshold: price = 0.90" and "elif inventory level < threshold: price = 1.10" is used. Specifically, the server updates price information via an SQL query and sends the new price information to the user's device.

[0805] (Application example 1)

[0806] 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."

[0807] Conventional demand forecasting systems require a lot of time and resources to compile data and make forecasts, making it difficult to respond quickly to fluctuations in demand. Furthermore, there are limited ways to present forecast results in a visually understandable manner, making it difficult for managers and store managers to make appropriate decisions. Furthermore, when it comes to dynamic pricing, there is a lack of functionality to automatically present effective strategies based on demand forecasts.

[0808] 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.

[0809] In this invention, the server includes means for collecting past data for demand forecasting, means for preprocessing the collected data, means for training a machine learning model based on the preprocessed data, means for forecasting future demand using the trained machine learning model, means for providing the forecast results to a user, means for dynamically setting prices based on the demand forecast results, means for predicting sales trends and proposing appropriate price increases or decreases, and means for visualizing the forecast results, thereby enabling store managers to quickly and effectively perform demand forecasting and optimize dynamic pricing and inventory management.

[0810] "Demand forecasting" is the prediction of future customer demand and sales volume based on past data.

[0811] "Data collection" refers to the collection of historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and related systems.

[0812] "Data preprocessing" refers to processing collected raw data, such as filling in missing data, correcting or deleting error values ​​and outliers, converting numerical data to a standard scale, and integrating different data.

[0813] A "machine learning model" is an algorithm that is trained to perform tasks such as prediction or classification using collected and preprocessed data as input.

[0814] "Dynamic pricing" refers to adjusting the price of goods and services in real time in response to changes in supply and demand.

[0815] "Collected Data" refers to historical sales data, consumer behavior data, seasonal data, and market trend data obtained from retailers and related systems via APIs or other means.

[0816] A "machine learning algorithm" is a method or procedure for performing machine learning tasks such as demand forecasting and classification based on various data, for example, LSTM, Random Forest, or XGBoost.

[0817] "Sales trend forecasting" refers to predicting future sales figures and sales trends based on past sales data and other related information.

[0818] A "price increase or decrease proposal" is a specific recommendation to increase or decrease the price of a product based on projected demand and inventory availability.

[0819] "Visualization of forecast results" refers to displaying predicted demand and sales trends in a visual format such as graphs and charts.

[0820] This invention relates to a demand forecasting and dynamic pricing system that utilizes AI and machine learning to predict future demand based on a wide range of data, including past sales data, and dynamically sets prices for products and services based on the results.

[0821] First, the server collects past sales data, consumer behavior data, seasonal data, and market trend data from retailers and related systems via API. For example, monthly sales data from the last fiscal year is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API. The data collected in this way is stored on the server as raw data.

[0822] Next, the server preprocesses the collected data. Missing data is imputed with the average value, and errors and outliers are detected and corrected or removed. Collected data from different data sources is integrated, converted to a standard scale (e.g., a range from 0 to 1), and organized into a unified dataset. The main software used in this process includes Python's Pandas and Numpy.

[0823] The server trains a machine learning model based on the preprocessed data. The dataset is divided into training data and test data, and a machine learning algorithm such as random forest or LSTM is selected to build the model. For example, past sales data and consumer behavior data are input to train the LSTM network. Libraries such as Scikit-learn and TensorFlow are used in this process.

[0824] The server uses the trained machine learning model to predict future demand. Test data can be input into the model to predict future sales trends. The prediction results are sent to the user's device and displayed in a visualized format, such as a line graph or bar chart. The results are presented in an intuitive and easy-to-understand format for managers and store administrators.

[0825] The server also has a means for dynamically setting prices based on the results of demand forecasts. It automatically adjusts prices by raising them during periods of increased demand and lowering them when there is excess inventory. Prices are dynamically set in conjunction with current inventory levels, promoting sales. This algorithm is reflected in the product database, and updated price information is sent to users' devices in real time.

[0826] For example, when weekly sales data, consumer behavior data, and market trend data are input, the system will suggest appropriate price increases or decreases based on the predicted sales trends for the following month.An example of a prompt to be input into the generative AI model is, "Based on last week's sales data, consumer behavior data, and market trend data, please suggest predicted sales trends and dynamic pricing for the next two weeks."

[0827] In this way, the present invention provides an efficient tool for predicting future sales trends and dynamic pricing in brick-and-mortar stores.

[0828] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0829] Step 1: Data collection

[0830] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and related systems via APIs. This may include, for example, retrieving monthly sales data from an SQL database or collecting customer web browsing history through the Google Analytics API. The input of this step is the data source, and the output is the collected raw data.

[0831] Step 2: Data Preprocessing

[0832] The server preprocesses the collected raw data by first imputing missing data with the mean value, then detecting and correcting or removing erroneous values ​​and outliers, normalizing numerical data to a range of 0 to 1, and integrating data collected from different data sources. The input of this step is the collected raw data, and the output is a preprocessed dataset.

[0833] Step 3: Train the model

[0834] The server uses the preprocessed data to train a machine learning model. First, it splits the dataset into training data and test data. Then it selects an algorithm, such as random forest or LSTM, and builds a model using the training data. The input of this step is the preprocessed data, and the output is a trained machine learning model.

[0835] Step 4: Demand forecast

[0836] The server uses the trained machine learning model to predict future demand. Specifically, it inputs test data into the model and calculates future sales trends. The inputs for this step are the test data and the trained model, and the output is predicted demand data.

[0837] Step 5: Providing prediction results

[0838] The server provides the forecast results to the user. For example, it displays the forecasted demand data on the user's terminal in a visual format such as a line graph or bar graph. The input of this step is the forecasted demand data, and the output is the visualized forecast results.

[0839] Step 6: Dynamic Pricing

[0840] The server dynamically sets prices based on the demand forecast results. Specifically, it raises product prices before the period when demand is predicted to increase, and lowers prices when there is excess inventory. It also takes into account current inventory levels and raises prices when inventory is low. The inputs to this step are predicted demand data and inventory data, and the output is dynamically set price information.

[0841] Step 7: Sales trend forecast and proposals

[0842] The server proposes price increases or decreases based on future sales trends. For example, by inputting weekly sales data, consumer behavior data, and market trend data, the server proposes appropriate price adjustments based on the predicted sales trends for the next month. The inputs for this step are the collected data and predicted sales trends, and the output is a specific price adjustment proposal.

[0843] Step 8: Visualize the prediction results

[0844] The server visualizes the predicted demand and sales trends. It displays the forecast results in line graphs and bar charts, and communicates them in an intuitive and easy-to-understand format to the user. The input to this step is the predicted demand data and sales trend data, and the output is visualized graphs and charts.

[0845] 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.

[0846] This invention combines a system that uses AI and machine learning to predict future sales trends and demand with an emotion engine that recognizes user emotions. The specific processing of the program used in this system is shown below.

[0847] 1. Data Collection

[0848] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and other related systems via APIs. For example, monthly sales data from the last year is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API.

[0849] 2. Data Preprocessing

[0850] The server imputes missing data from the collected data, for example by imputing the average value or by using values ​​before and after the data, and then detects erroneous or outlier values ​​and corrects or deletes them as necessary.

[0851] The server converts the numerical data to a standard scale (e.g., a range from 0 to 1) to enable comparisons between different data sets, and combines and formats data collected from different sources into a single integrated dataset.

[0852] 3. Model Training

[0853] The server splits the formatted dataset into training data and test data, typically using 80% of the data as training data and 20% as test data.

[0854] The server selects and initializes the machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) that is best suited for demand forecasting. For example, it selects the LSTM model and sets the necessary hyperparameters.

[0855] The server trains the model using the selected machine learning algorithm by providing training data. Specifically, it inputs past sales data into an LSTM network and builds a model to predict future sales.

[0856] 4. Demand forecasting

[0857] The server inputs test data using the trained model to predict future sales trends and demand, for example, forecasting sales for the next month, and outputs the results as data.

[0858] 5. Outputting the results

[0859] The server then converts the forecast results into an appropriate format and sends them to the user's device. For example, the sales forecast results are visualized as line graphs or bar charts and displayed on the user's dashboard.

[0860] The emotion engine uses a camera and microphone to analyze the user's facial expressions and voice when presenting results, and dynamically changes the way prediction results are presented based on the recognized emotions.

[0861] For example: If the user is surprised, show them some additional information to help them calm down. If the user is confused, show them a detailed explanation.

[0862] 6. Dynamic Pricing

[0863] The server runs an algorithm that dynamically sets product prices based on predicted demand, for example, raising prices by 10% ahead of periods of predicted increased demand and lowering prices during periods of overstock.

[0864] The server takes into account the current inventory level and optimizes pricing, for example lowering the price if inventory is excessive and raising the price if inventory is low.

[0865] The server reflects the optimized pricing in the product database and transmits the new price information to the user's terminal in real time.

[0866] In addition to the demand forecasting and dynamic pricing functions provided by this system, by integrating user emotion recognition with an emotion engine, it is possible to provide a more intuitive and adaptive interface for users, which will enable companies to improve not only inventory management and production scheduling but also the user experience.

[0867] The processing flow will be explained below.

[0868] Step 1:

[0869] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and other related systems via APIs. For example, monthly sales data from the last year is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API.

[0870] Step 2:

[0871] The server imputes missing data from the collected data, for example by imputing the average value or by using values ​​before and after the data, and then detects erroneous or outlier values ​​and corrects or deletes them as necessary.

[0872] Step 3:

[0873] The server converts numerical data to a standard scale (e.g., a range from 0 to 1) to allow comparisons between different data sets, and combines and formats data collected from different sources into a single integrated dataset.

[0874] Step 4:

[0875] The server splits the formatted dataset into training data and test data, typically using 80% of the data as training data and 20% as test data.

[0876] Step 5:

[0877] The server selects and initializes the machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) that is best suited for demand forecasting. For example, it selects the LSTM model and sets the necessary hyperparameters.

[0878] Step 6:

[0879] The server trains the model using the selected machine learning algorithm by providing training data. Specifically, it inputs past sales data into an LSTM network and builds a model to predict future sales.

[0880] Step 7:

[0881] The server inputs test data using the trained model to predict future sales trends and demand, for example, forecasting sales for the next month, and outputs the results as data.

[0882] Step 8:

[0883] The server then converts the forecast results into an appropriate format and sends them to the user's device. For example, the sales forecast results are visualized as line graphs or bar charts and displayed on the user's dashboard.

[0884] Step 9:

[0885] The device uses a camera and microphone to collect the user's facial expressions and voice data, which the emotion engine analyzes to recognize the user's emotions, such as whether the user is confused or surprised.

[0886] Step 10:

[0887] The server dynamically changes how the predictions are presented based on the recognized emotion: for example, if the user is surprised, it may display additional information to calm them down, or if the user is confused, it may display additional detailed explanations or instructions.

[0888] Step 11:

[0889] The server runs an algorithm that dynamically sets product prices based on predicted demand, for example, raising prices by 10% ahead of periods of predicted increased demand and lowering prices during periods of overstock.

[0890] Step 12:

[0891] The server takes into account the current inventory level and optimizes pricing, for example lowering the price if inventory is excessive and raising the price if inventory is low.

[0892] Step 13:

[0893] The server reflects the optimized pricing in the product database and transmits the new price information to the user's terminal in real time.

[0894] Step 14:

[0895] Users can check new price information and forecast results through their devices, and can adjust their inventory management and sales strategies accordingly.

[0896] Through this series of steps, the system of the present invention achieves highly accurate demand forecasting and dynamic pricing, and also provides an adaptive interface that responds to user emotions, allowing companies to optimize all aspects of inventory management, product production, and sales strategies.

[0897] Example 2

[0898] 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."

[0899] Conventional demand forecasting systems use past data to predict future demand, but they are unable to provide an interface that takes user emotions into account, resulting in a poor user experience. In particular, because the user's psychological state is not reflected when presenting the forecast results, there is a lack of support for the user to understand the results and take action. Conventional systems also lack the ability to dynamically set prices.

[0900] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0901] In this invention, the server includes means for collecting past data for demand forecasting, means for preprocessing the collected data, means for training a machine learning model based on the preprocessed data, means for forecasting future demand using the trained machine learning model, means for providing the forecast results to a user, means for recognizing the user's emotions, means for dynamically presenting the forecast results based on the user's emotions, and means for dynamically setting prices based on the demand forecast results. This makes it possible to provide an interface that reflects the user's emotions, improving the user experience and providing more support for understanding and taking action on the demand forecast results. Furthermore, dynamic pricing enables flexible price management in response to market fluctuations.

[0902] "Demand forecasting" is the process of predicting future consumer demand based on past and current data.

[0903] "Data Collection Measures" refers to the ability to collect necessary data from retailers and other related systems.

[0904] "Data preprocessing means" refers to functions that fill in missing values ​​in collected data, correct or delete error values ​​or outliers, and perform scaling and integration.

[0905] "Machine learning model training means" refers to a function that uses preprocessed data to train a machine learning algorithm and create a predictive model.

[0906] "Demand forecasting means" refers to the ability to forecast future demand using a trained machine learning model.

[0907] "Means of providing to users" refers to the function of providing prediction results to users in an appropriate format and presenting them in an easy-to-understand manner.

[0908] "Emotion recognition means" refers to a function that analyzes the user's facial expressions and voice to recognize the user's emotional state.

[0909] "Dynamic presentation means" refers to a function that dynamically changes the way prediction results are presented based on the recognized user emotions.

[0910] "Dynamic pricing means" refers to a function that dynamically sets product prices in real time based on demand forecast results.

[0911] This invention combines a system that uses AI and machine learning to predict future sales trends and demand with an emotion engine that recognizes user emotions. The specific processing of the program used in this system is described in detail below.

[0912] First, the server collects data. It collects past sales data, consumer behavior data, seasonal data, market trend data, etc. from retailers and related systems. Specifically, it retrieves last year's monthly sales data from an SQL database and collects users' web browsing history using the Google Analytics API. The data collected in this way includes information from a variety of sources.

[0913] Next, the server performs data preprocessing. Missing data is imputed from the collected data, and errors and outliers are detected and corrected or removed. Specific processing methods include average imputation and imputation based on surrounding values. All numerical data is also converted to a standard scale to enable comparisons between different data sets. Data collected from different sources is combined into a single integrated dataset and formatted.

[0914] The server then performs model training. The formatted dataset is divided into training data and test data, and the machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) optimal for demand forecasting is selected and initialized. For example, an LSTM model is selected and the necessary hyperparameters are set. Training data is input using this selected algorithm to train the model. Specifically, past sales data is input into the LSTM network to build a model for predicting future sales.

[0915] Next, the server performs demand forecasting. It uses the trained model to input test data and predict future sales trends and demand. For example, it predicts sales for the next month and outputs the results as data.

[0916] The prediction results are provided from the server to the user's device in an appropriate format. The prediction results are visualized in line graphs and bar charts and displayed on the user's dashboard. The emotion engine uses a camera and microphone to analyze the user's facial expressions and voice when presenting the results, recognizing the user's emotions. Based on the recognized emotions, the way the prediction results are presented is dynamically changed. For example, if the user is surprised, additional information to calm them down is displayed, and if the user is confused, a detailed explanation is displayed.

[0917] Dynamic pricing is also an important feature of this system. The server runs an algorithm that dynamically sets product prices based on predicted demand. For example, the price can be raised by 10% before a period when demand is predicted to increase, and lowered when there is excess inventory. Furthermore, pricing is optimized taking into account current inventory levels. For example, if there is excess inventory, the price is lowered, and if inventory is low, the price is raised. The optimized pricing is reflected in the product database, and new price information is sent to users' devices in real time.

[0918] The system integrates demand forecasting and dynamic pricing functions, as well as user emotion recognition through an emotion engine, to provide a more intuitive and adaptive interface for users, helping companies not only to optimize inventory management and production schedules, but also to improve the user experience.

[0919] Specific examples

[0920] For example, the server retrieves last year's sales data from an SQL database and collects users' web browsing history using the Google Analytics API. It then uses mean imputation to impute missing data and scales the data to a range of 0 to 1 using MinMaxScaler. It selects an LSTM model and trains it using the training data with the fit() function. It uses the predict() function to predict next month's sales and displays the results in a line graph. The emotion engine analyzes emotions from the user's facial expressions and voice and dynamically changes the way the prediction results are presented. Based on demand, the server may raise prices by 10% or optimize prices, update the product database, and notify the user's device of the new price information.

[0921] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0922] Program processing flow and specific explanation

[0923] Step 1: Data collection

[0924] Input: Database connection information for retailers and related systems, API keys

[0925] Specific behavior:

[0926] The server uses an SQL query to retrieve last year's monthly sales data from the database.

[0927] The server uses the Google Analytics API to collect the user's web browsing history for the past six months.

[0928] Output: Collected dataset of historical sales data and consumer behavior data

[0929] Step 2: Data Preprocessing

[0930] Input: raw data collected, missing data, data containing error values ​​and outliers

[0931] Specific behavior:

[0932] The server generates a data frame using the Pandas library and performs imputation of missing values ​​(e.g., mean imputation).

[0933] The server runs an outlier detection algorithm to detect and remove or correct outliers.

[0934] The server applies MinMaxScaler to scale the numeric data to the range 0 to 1.

[0935] The server combines and formats the data collected from different sources into a single integrated data set.

[0936] Output: Preprocessed consolidated dataset

[0937] Step 3: Model training

[0938] Input: Preprocessed unified dataset, dataset split into training and test data

[0939] Specific behavior:

[0940] The server splits the combined dataset into training data (80%) and test data (20%) using the Scikit-learn or Keras libraries.

[0941] The server selects an LSTM model and sets hyperparameters (e.g., learning rate, number of epochs).

[0942] The server uses the training data to train the LSTM model using the fit() function.

[0943] Output: A trained LSTM model

[0944] Step 4: Demand forecast

[0945] Input: trained model, test data

[0946] Specific behavior:

[0947] The server uses the predict() function to forecast future sales trends and demand based on the test data.

[0948] Specifically, it generates sales forecast results for the next month.

[0949] Output: Future sales forecast results

[0950] Step 5: Output the results

[0951] Input: Future sales forecast results

[0952] Specific behavior:

[0953] The server uses the Matplotlib and Seaborn libraries to visualize the forecast results in line and bar graphs.

[0954] The server sends the visualized data to the user's device and displays it on a dashboard.

[0955] Output: Visualized data for dashboard display

[0956] Step 6: Emotion Recognition

[0957] Input: User's facial expressions, voice data

[0958] Specific behavior:

[0959] The emotion engine analyzes facial and voice data collected from the camera and microphone to recognize the user's emotional state.

[0960] The emotion engine uses facial recognition and voice analysis algorithms.

[0961] Output: Recognized emotional state of the user

[0962] Step 7: Dynamic Presentation

[0963] Input: Recognized user emotional state, future sales forecast results

[0964] Specific behavior:

[0965] The server dynamically changes the way predictions are presented based on the user's perceived emotions: if the user is surprised, it displays additional reassuring information, and if the user is confused, it displays a detailed explanation.

[0966] Output: Dynamically changed presentation results based on user emotion

[0967] Step 8: Dynamic Pricing

[0968] Input: Future sales forecast results, current inventory amount

[0969] Specific behavior:

[0970] The server dynamically sets product prices based on predicted demand, for example by raising prices by 10% before periods when demand is predicted to increase, and lowering prices when there is excess inventory.

[0971] The server optimizes pricing by taking into account inventory levels. Specifically, if inventory is excessive, the price is reduced by 5%, and if inventory is low, the price is increased by 5%.

[0972] Output: Optimized pricing information

[0973] Step 9: Update and submit pricing information

[0974] Input: Optimized Pricing Information

[0975] Specific behavior:

[0976] The server updates the product database and transmits new price information to the user terminal in real time.

[0977] Output: Updated price information, real-time notification to user terminal

[0978] (Application example 2)

[0979] 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."

[0980] Conventional demand forecasting systems primarily use historical data and machine learning to predict future sales trends, but they are unable to consider the user's emotional state. As a result, it is difficult to provide optimal interfaces and adaptive feedback for users. In particular, the lack of emotion recognition makes it difficult for users to accept and understand the prediction results, resulting in reduced usability. Furthermore, while dynamic pricing is automatically set based on demand forecasts, it fails to sufficiently increase user satisfaction because pricing based on user emotions and behavior is not taken into account.

[0981] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0982] In this invention, the server includes means for collecting past data for demand forecasting, means for preprocessing the collected data, means for training a machine learning model based on the preprocessed data, means for forecasting future demand using the trained machine learning model, means for providing the forecast results to a user, means for dynamically setting prices based on the demand forecast results, and means for recognizing the user's emotions and dynamically changing the presentation method of the forecast results and the interface based on the emotions. This makes it possible to provide more intuitive and adaptive feedback while taking the user's emotions into consideration. Dynamic pricing based on the user's emotions is also realized, thereby improving user satisfaction.

[0983] "Demand forecasting" refers to analyzing past data and predicting future sales trends and demand.

[0984] "Means for collection" refers to a function that allows necessary data to be acquired from various data sources.

[0985] "Means for data preprocessing" refers to the function of completing missing data and correcting outliers in collected data, and processing it to prepare it in a format that can be analyzed.

[0986] A "machine learning model" is a trained algorithm used to predict future demand based on historical data.

[0987] "Training" refers to the process of inputting data into a machine learning model and adjusting the model's parameters to improve its predictive accuracy.

[0988] A "demand forecasting means" is one that has the ability to forecast future sales trends and demand using trained machine learning models.

[0989] "Means for providing to the user" refers to a function that provides an interface for displaying prediction results to the user.

[0990] "Dynamic pricing" refers to the ability to adjust product prices in real time based on predicted demand.

[0991] "Means for recognizing user emotions" refers to technology that uses a camera and microphone to analyze the user's facial expressions and voice to identify their emotional state.

[0992] "Means for dynamically changing the interface presentation method" refers to a function that adaptively changes the screen display and information presentation method based on the user's emotions.

[0993] A system for implementing the present invention is configured as follows.

[0994] First, the server collects historical data for demand forecasting. The collected data includes past sales data, consumer behavior data, seasonal data, and market trend data. This data is collected by querying an SQL database or retrieving data through APIs.

[0995] The server then preprocesses the collected data, which includes imputing missing data, correcting errors and outliers, converting numerical data to a standard scale, and integrating data from different sources. This process prepares the data in a format that can be analyzed.

[0996] The server then trains a machine learning model based on the preprocessed data. The machine learning model to be used may be LSTM, Random Forest, XGBoost, or other models. The data is separated into training data and test data, and the model is trained using the selected algorithm. This process builds a highly accurate model for predicting future demand.

[0997] The server uses the trained model to predict future demand, and the predicted demand data is provided to the user. For example, the sales forecast results are displayed on the user's device as a line graph or bar chart.

[0998] The server also uses a camera and microphone to recognize the user's emotions. Using audio and image analysis software such as EmotionEngine, the server analyzes the user's facial expressions and voice to identify their emotions. Based on this identified emotion, the server dynamically changes the way prediction results are presented and the interface. For example, if the user is surprised, the server displays additional information along with the prediction result.

[0999] The server also dynamically sets prices based on the results of demand forecasts. Prices are raised if demand is predicted to increase, and lowered if there is excess inventory. This allows companies to optimize inventory management and pricing.

[1000] This system provides the following advantages: Improved usability, as users can intuitively understand the information they want; and Dynamic pricing, which maximizes profits for businesses. This is a specific embodiment of the present invention.

[1001] For example, if the user is surprised, a message will appear on the screen saying, "Don't worry. We'll show you more information." Additionally, if the current stock level is low, the product price will automatically increase.

[1002] Example prompt for generative AI model:

[1003] "How can I forecast sales trends and demand and create a dynamic interface that responds to user emotions?"

[1004] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1005] Step 1:

[1006] The server collects historical data for demand forecasting, including sales data from an SQL database, consumer behavior data, seasonality data, and market trend data through a web API. The input is data provided by various data sources, and the output is the collected raw data set.

[1007] Step 2:

[1008] The server preprocesses the collected data, which includes imputing missing data, correcting errors and outliers, and converting numerical data to a standard scale (ranging from 0 to 1). It also integrates data from different data sources. The input is the raw dataset collected in step 1, and the output is the preprocessed dataset.

[1009] Step 3:

[1010] The server trains a machine learning model based on the preprocessed data. First, it splits the data into training data and test data. Then it trains the model using a selected machine learning algorithm (e.g., LSTM, Random Forest, etc.). The input is the preprocessed dataset, and the output is the trained machine learning model.

[1011] Step 4:

[1012] The server uses the trained model to predict future demand. Test data is input to the trained model to predict future sales trends and demand. The inputs are the test data and the trained model, and the output is the predicted result data.

[1013] Step 5:

[1014] The server provides the prediction results to the user. The prediction results are visualized as line graphs or bar graphs and displayed on the user's device. The input is the prediction result data, and the output is the visual data displayed on the user's device.

[1015] Step 6:

[1016] The server uses a camera and microphone to recognize the user's emotions. It uses the EmotionEngine to analyze the user's facial expressions and voice to identify the user's emotions. The input is image and audio data obtained from the camera and microphone, and the output is the recognized emotion data.

[1017] Step 7:

[1018] The server dynamically changes the interface presentation based on the recognized emotion, for example, displaying additional information if the user is surprised, or a detailed explanation if the user is confused. The input is the recognized emotion data and the predicted result data, and the output is the adjusted interface.

[1019] Step 8:

[1020] The server dynamically sets prices based on the demand forecast results. When the forecast demand is high, the price is raised, and when there is excess inventory, the price is lowered. The input is the forecast result data and current inventory data, and the output is optimized price information.

[1021] Step 9:

[1022] The server sends new price information to the user's terminal in real time. The input is the optimized price information, and the output is the new price information displayed on the user's terminal.

[1023] 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.

[1024] 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.

[1025] 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.

[1026] [Fourth embodiment]

[1027] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1028] 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.

[1029] 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).

[1030] 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.

[1031] 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.

[1032] 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).

[1033] 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.

[1034] 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.

[1035] 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.

[1036] 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.

[1037] 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.

[1038] 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.

[1039] 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."

[1040] This invention provides a system that uses AI and machine learning to predict future sales trends and demand. The specific processing of the program used in this system is shown below.

[1041] 1. Data Collection

[1042] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and other related systems via APIs.

[1043] Example: Last year's monthly sales data is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API.

[1044] 2. Data Preprocessing

[1045] The server imputes missing data from the collected raw data, for example, by imputing the average value or by connecting previous and next values.

[1046] The server detects erroneous or outlier values ​​and corrects or removes them as necessary.

[1047] The server converts the numeric data to a standard scale (e.g., a range from 0 to 1) to allow comparison between different data.

[1048] The server combines and formats the data collected from different sources into a single integrated data set.

[1049] 3. Model Training

[1050] The server splits the dataset into training data and test data, typically 80% as training data and 20% as test data.

[1051] The server selects and initializes a machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) suitable for demand forecasting.

[1052] The server uses the selected machine learning algorithm to build a model using the training data, for example, inputting the data into an LSTM network to train a demand forecasting model.

[1053] 4. Demand forecasting

[1054] The server inputs test data using the trained machine learning model to predict future sales trends and demand.

[1055] Example: Use test data to forecast sales for the next month and check the results.

[1056] 5. Outputting the results

[1057] The server formats the prediction results appropriately and sends them to the user's device.

[1058] Example: Sales forecast results are visualized as line graphs and bar graphs and displayed on the user's dashboard.

[1059] 6. Dynamic Pricing

[1060] A server runs an algorithm that dynamically sets prices for products based on demand forecasts.

[1061] For example, increase prices by 10% before a period when demand is predicted to increase, or lower prices during periods of excess inventory to encourage sales.

[1062] The server takes into account the current stock amount and performs processing to lower the price if there is excess stock and to raise the price if there is low stock.

[1063] The server reflects the optimized pricing in the product database and transmits the new price information to the user's terminal.

[1064] The system's forecasting and pricing capabilities enable users to highly optimize inventory management, production scheduling, and dynamic pricing, thereby enabling companies to make more effective decisions in sales planning and maximize profits.

[1065] The processing flow will be explained below.

[1066] Step 1:

[1067] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and other related systems via APIs. For example, monthly sales data from the last year is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API.

[1068] Step 2:

[1069] The server imputes missing data from the collected data, for example by imputing the average value or by using values ​​before and after the data, and then detects erroneous or outlier values ​​and corrects or deletes them as necessary.

[1070] Step 3:

[1071] The server converts the numerical data to a standard scale (e.g., a range from 0 to 1) to enable comparisons between different data sets, and combines and formats data collected from different sources into a single integrated dataset.

[1072] Step 4:

[1073] The server splits the formatted dataset into training data and test data, typically using 80% of the data as training data and 20% as test data.

[1074] Step 5:

[1075] The server selects and initializes the machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) that is best suited for demand forecasting. For example, it selects the LSTM model and sets the necessary hyperparameters.

[1076] Step 6:

[1077] The server trains the model using the selected machine learning algorithm by providing training data. Specifically, it inputs past sales data into an LSTM network and builds a model to predict future sales.

[1078] Step 7:

[1079] The server inputs test data using the trained model to predict future sales trends and demand, for example, forecasting sales for the next month, and outputs the results as data.

[1080] Step 8:

[1081] The server then converts the forecast results into an appropriate format and sends them to the user's device. For example, the sales forecast results are visualized as line graphs or bar charts and displayed on the user's dashboard.

[1082] Step 9:

[1083] The server runs an algorithm that dynamically sets product prices based on predicted demand, for example, raising prices by 10% ahead of periods of predicted increased demand and lowering prices during periods of overstock.

[1084] Step 10:

[1085] The server takes into account the current inventory level and optimizes pricing, for example lowering the price if inventory is excessive and raising the price if inventory is low.

[1086] Step 11:

[1087] The server reflects the optimized pricing in the product database and transmits the new price information to the user's terminal in real time.

[1088] Through this series of steps, the system of the present invention can achieve highly accurate demand forecasting and dynamic pricing, optimizing a company's inventory management and sales planning.

[1089] Example 1

[1090] 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."

[1091] In recent years, retailers have found it difficult to accurately forecast sales trends and demand, which has led to problems in inventory management and pricing. In particular, the processes of data collection, data preprocessing, model training, and dynamic pricing required to predict demand fluctuations are complex, and a system for managing these processes efficiently and effectively is needed.

[1092] 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.

[1093] In this invention, the server includes: means for collecting past data from retailers and related systems; means for preprocessing the collected data by imputing missing data, detecting and removing outliers, standard scale conversion, and data combination; means for dividing the preprocessed data into training data and test data and selecting and training a machine learning algorithm suitable for demand forecasting; means for predicting future sales trends and demand using the trained machine learning model; means for visually displaying the prediction results and sending them to a user's terminal; and means for dynamically setting product prices based on the demand forecast results and inventory levels and updating the product database. This enables accurate prediction of demand fluctuations and efficient and effective inventory management and pricing.

[1094] "Data collection" is the process of obtaining historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and related systems.

[1095] "Data preprocessing" is the process of filling in missing data from collected raw data, detecting and removing outliers, and converting the data into a standard scale to put the data into a unified format.

[1096] "Machine learning algorithms" are mathematical models or methods used to predict future trends or demand based on a dataset. This invention uses LSTM, Random Forest, and XGBoost as examples.

[1097] "Training data" refers to the large dataset used to build and train machine learning models.

[1098] "Test data" refers to a dataset used to evaluate the performance of a trained machine learning model.

[1099] "Demand forecasting" is the process of using trained machine learning models to predict future sales trends and product demand.

[1100] "Dynamic pricing" is the process of automatically adjusting product prices based on demand forecasts and inventory levels.

[1101] "Product database" means a digital database for managing and storing product and pricing information.

[1102] "Visual display" is the process of presenting the results of prediction to a user in a visually easy-to-understand format such as a graph or table.

[1103] "User terminal" refers to a device, such as a PC or smartphone, that receives and displays prediction results and price information.

[1104] This invention provides a system that uses AI and machine learning to predict future sales trends and demand. This system operates in cooperation with a server, terminals, and users.

[1105] Data collection

[1106] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and related systems via APIs, specifically retrieving sales data from an SQL database and using the Google Analytics API to pull web browsing history data.

[1107] Example: Retrieve last year's monthly sales data using an SQL query and collect user web browsing history using the Google Analytics API.

[1108] Data Preprocessing

[1109] The server performs preprocessing on the collected raw data, including missing data imputation, outlier detection and removal, standard scale conversion, and data merging. Missing data is imputed using mean imputation, and outliers are filtered based on standard deviation. Data is standardized to a range of 0 to 1, and different data sources are combined into an integrated dataset.

[1110] Example: Use the Pandas library to impute NaN values ​​in a DataFrame, remove outliers, and perform normalization.

[1111] Model learning

[1112] The server splits the preprocessed data into training data and test data, selects and initializes a machine learning algorithm such as LSTM, Random Forest, or XGBoost, and builds and trains a model based on the training data.

[1113] Example: Using Sci-kit Learn, split the dataset into 80% training data and 20% test data, and build and train an LSTM model.

[1114] Demand forecasting

[1115] The server uses the trained model to predict future sales trends and demand based on the test data, and the prediction results are saved in a format such as a CSV file.

[1116] Example: Use test data to forecast sales for the next month and save the results to a file.

[1117] Output of results

[1118] The server visually displays the prediction results and sends them to the user's device. For example, it uses Matplotlib or Seaborn to draw line graphs and bar charts and display them on a dashboard.

[1119] Example: Generate a graph based on the prediction results and display it on a dashboard.

[1120] Dynamic Pricing

[1121] The server dynamically sets product prices based on demand forecasts and inventory levels, and updates the product database with this information. Based on a pricing algorithm, prices are increased during periods of increased demand and decreased when inventory levels are excessive.

[1122] Example: Raise prices by 10% when demand is predicted to increase, or lower prices when there is excess inventory.

[1123] Example prompt: "Use AI and machine learning to generate a Python script that uses the past year's sales data to forecast next month's sales and dynamically set pricing. Please include any libraries needed and specific code examples."

[1124] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1125] Step 1: Data collection

[1126] The server collects historical sales data, consumer behavior data, seasonal data, and market trend data from retailers and related systems. It queries an SQL database and retrieves web browsing history data using the Google Analytics API. The input is sales data, web click data, etc., and the output is a consolidated dataset containing these data. Specifically, the server securely collects data using API keys and authentication tokens.

[1127] Step 2: Data Preprocessing

[1128] The server preprocesses the collected data, including imputing missing data, detecting and removing outliers, converting to a standard scale, and combining different data sources. The input is raw data, and the output is preprocessed, clean data. The server uses the Pandas library to process the missing data, such as "df['Sales'].fillna(df['Sales'].mean(), inplace=True)", to filter outliers and convert to a standard scale. It also combines data from different data sources to generate a unified dataset.

[1129] Step 3: Split the data

[1130] The server splits the preprocessed data into training data and test data. The input is the clean dataset, and the output is the training data and test data. The server uses Sci-kit Learn and executes code such as "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)".

[1131] Step 4: Model training

[1132] The server uses the training data to train a machine learning model. The input is the training data, and the output is the trained model. The server selects an algorithm such as LSTM, Random Forest, or XGBoost, initializes it with code such as "from keras.models import Sequential" and "model = Sequential()", and trains the model with code such as "model.fit(X_train, y_train, epochs=50, batch_size=32, validation_split=0.2)".

[1133] Step 5: Demand forecast

[1134] The server uses the trained model to predict future sales trends and demand based on the test data. The input is the test data and the output is the prediction result. The server makes the prediction using "y_pred = model.predict(X_test)" and saves the prediction result in a format such as a CSV file.

[1135] Step 6: Output the results

[1136] The server visually displays the prediction results and sends them to the user's device. The input is the prediction results, and the output is the visually displayed data. The server uses Matplotlib and Seaborn to draw graphs, visualizes them with code such as "plt.plot(prediction result)", and displays them on a dashboard.

[1137] Step 7: Dynamic Pricing

[1138] The server dynamically sets product prices based on demand forecast results and inventory levels, and reflects this information in the product database. The inputs are forecast results and inventory data, and the output is dynamically set price information. The server adjusts prices during periods of increased demand, such as "item_price = 1.10", and updates the new price information in the product database. Conditional branching such as "if inventory level > threshold: price = 0.90" and "elif inventory level < threshold: price = 1.10" is used. Specifically, the server updates price information via an SQL query and sends the new price information to the user's device.

[1139] (Application example 1)

[1140] 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."

[1141] Conventional demand forecasting systems require a lot of time and resources to compile data and make forecasts, making it difficult to respond quickly to fluctuations in demand. Furthermore, there are limited ways to present forecast results in a visually understandable manner, making it difficult for managers and store managers to make appropriate decisions. Furthermore, when it comes to dynamic pricing, there is a lack of functionality to automatically present effective strategies based on demand forecasts.

[1142] 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.

[1143] In this invention, the server includes means for collecting past data for demand forecasting, means for preprocessing the collected data, means for training a machine learning model based on the preprocessed data, means for forecasting future demand using the trained machine learning model, means for providing the forecast results to a user, means for dynamically setting prices based on the demand forecast results, means for predicting sales trends and proposing appropriate price increases or decreases, and means for visualizing the forecast results, thereby enabling store managers to quickly and effectively perform demand forecasting and optimize dynamic pricing and inventory management.

[1144] "Demand forecasting" is the prediction of future customer demand and sales volume based on past data.

[1145] "Data collection" refers to the collection of historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and related systems.

[1146] "Data preprocessing" refers to processing collected raw data, such as filling in missing data, correcting or deleting error values ​​and outliers, converting numerical data to a standard scale, and integrating different data.

[1147] A "machine learning model" is an algorithm that is trained to perform tasks such as prediction or classification using collected and preprocessed data as input.

[1148] "Dynamic pricing" refers to adjusting the price of goods and services in real time in response to changes in supply and demand.

[1149] "Collected Data" refers to historical sales data, consumer behavior data, seasonal data, and market trend data obtained from retailers and related systems via APIs or other means.

[1150] A "machine learning algorithm" is a method or procedure for performing machine learning tasks such as demand forecasting and classification based on various data, for example, LSTM, Random Forest, or XGBoost.

[1151] "Sales trend forecasting" refers to predicting future sales figures and sales trends based on past sales data and other related information.

[1152] A "price increase or decrease proposal" is a specific recommendation to increase or decrease the price of a product based on projected demand and inventory availability.

[1153] "Visualization of forecast results" refers to displaying predicted demand and sales trends in a visual format such as graphs and charts.

[1154] This invention relates to a demand forecasting and dynamic pricing system that utilizes AI and machine learning to predict future demand based on a wide range of data, including past sales data, and dynamically sets prices for products and services based on the results.

[1155] First, the server collects past sales data, consumer behavior data, seasonal data, and market trend data from retailers and related systems via API. For example, monthly sales data from the last fiscal year is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API. The data collected in this way is stored on the server as raw data.

[1156] Next, the server preprocesses the collected data. Missing data is imputed with the average value, and errors and outliers are detected and corrected or removed. Collected data from different data sources is integrated, converted to a standard scale (e.g., a range from 0 to 1), and organized into a unified dataset. The main software used in this process includes Python's Pandas and Numpy.

[1157] The server trains a machine learning model based on the preprocessed data. The dataset is divided into training data and test data, and a machine learning algorithm such as random forest or LSTM is selected to build the model. For example, past sales data and consumer behavior data are input to train the LSTM network. Libraries such as Scikit-learn and TensorFlow are used in this process.

[1158] The server uses the trained machine learning model to predict future demand. Test data can be input into the model to predict future sales trends. The prediction results are sent to the user's device and displayed in a visualized format, such as a line graph or bar chart. The results are presented in an intuitive and easy-to-understand format for managers and store administrators.

[1159] The server also has a means for dynamically setting prices based on the results of demand forecasts. It automatically adjusts prices by raising them during periods of increased demand and lowering them when there is excess inventory. Prices are dynamically set in conjunction with current inventory levels, promoting sales. This algorithm is reflected in the product database, and updated price information is sent to users' devices in real time.

[1160] For example, when weekly sales data, consumer behavior data, and market trend data are input, the system will suggest appropriate price increases or decreases based on the predicted sales trends for the following month.An example of a prompt to be input into the generative AI model is, "Based on last week's sales data, consumer behavior data, and market trend data, please suggest predicted sales trends and dynamic pricing for the next two weeks."

[1161] In this way, the present invention provides an efficient tool for predicting future sales trends and dynamic pricing in brick-and-mortar stores.

[1162] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1163] Step 1: Data collection

[1164] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and related systems via APIs. This may include, for example, retrieving monthly sales data from an SQL database or collecting customer web browsing history through the Google Analytics API. The input of this step is the data source, and the output is the collected raw data.

[1165] Step 2: Data Preprocessing

[1166] The server preprocesses the collected raw data by first imputing missing data with the mean value, then detecting and correcting or removing erroneous values ​​and outliers, normalizing numerical data to a range of 0 to 1, and integrating data collected from different data sources. The input of this step is the collected raw data, and the output is a preprocessed dataset.

[1167] Step 3: Train the model

[1168] The server uses the preprocessed data to train a machine learning model. First, it splits the dataset into training data and test data. Then it selects an algorithm, such as random forest or LSTM, and builds a model using the training data. The input of this step is the preprocessed data, and the output is a trained machine learning model.

[1169] Step 4: Demand forecast

[1170] The server uses the trained machine learning model to predict future demand. Specifically, it inputs test data into the model and calculates future sales trends. The inputs for this step are the test data and the trained model, and the output is predicted demand data.

[1171] Step 5: Providing prediction results

[1172] The server provides the forecast results to the user. For example, it displays the forecasted demand data on the user's terminal in a visual format such as a line graph or bar graph. The input of this step is the forecasted demand data, and the output is the visualized forecast results.

[1173] Step 6: Dynamic Pricing

[1174] The server dynamically sets prices based on the demand forecast results. Specifically, it raises product prices before the period when demand is predicted to increase, and lowers prices when there is excess inventory. It also takes into account current inventory levels and raises prices when inventory is low. The inputs to this step are predicted demand data and inventory data, and the output is dynamically set price information.

[1175] Step 7: Sales trend forecast and proposals

[1176] The server proposes price increases or decreases based on future sales trends. For example, by inputting weekly sales data, consumer behavior data, and market trend data, the server proposes appropriate price adjustments based on the predicted sales trends for the next month. The inputs for this step are the collected data and predicted sales trends, and the output is a specific price adjustment proposal.

[1177] Step 8: Visualize the prediction results

[1178] The server visualizes the predicted demand and sales trends. It displays the forecast results in line graphs and bar charts, and communicates them in an intuitive and easy-to-understand format to the user. The input to this step is the predicted demand data and sales trend data, and the output is visualized graphs and charts.

[1179] 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.

[1180] This invention combines a system that uses AI and machine learning to predict future sales trends and demand with an emotion engine that recognizes user emotions. The specific processing of the program used in this system is shown below.

[1181] 1. Data Collection

[1182] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and other related systems via APIs. For example, monthly sales data from the last year is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API.

[1183] 2. Data Preprocessing

[1184] The server imputes missing data from the collected data, for example by imputing the average value or by using values ​​before and after the data, and then detects erroneous or outlier values ​​and corrects or deletes them as necessary.

[1185] The server converts the numerical data to a standard scale (e.g., a range from 0 to 1) to enable comparisons between different data sets, and combines and formats data collected from different sources into a single integrated dataset.

[1186] 3. Model Training

[1187] The server splits the formatted dataset into training data and test data, typically using 80% of the data as training data and 20% as test data.

[1188] The server selects and initializes the machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) that is best suited for demand forecasting. For example, it selects the LSTM model and sets the necessary hyperparameters.

[1189] The server trains the model using the selected machine learning algorithm by providing training data. Specifically, it inputs past sales data into an LSTM network and builds a model to predict future sales.

[1190] 4. Demand forecasting

[1191] The server inputs test data using the trained model to predict future sales trends and demand, for example, forecasting sales for the next month, and outputs the results as data.

[1192] 5. Outputting the results

[1193] The server then converts the forecast results into an appropriate format and sends them to the user's device. For example, the sales forecast results are visualized as line graphs or bar charts and displayed on the user's dashboard.

[1194] The emotion engine uses a camera and microphone to analyze the user's facial expressions and voice when presenting results, and dynamically changes the way prediction results are presented based on the recognized emotions.

[1195] For example: If the user is surprised, show them some additional information to help them calm down. If the user is confused, show them a detailed explanation.

[1196] 6. Dynamic Pricing

[1197] The server runs an algorithm that dynamically sets product prices based on predicted demand, for example, raising prices by 10% ahead of periods of predicted increased demand and lowering prices during periods of overstock.

[1198] The server takes into account the current inventory level and optimizes pricing, for example lowering the price if inventory is excessive and raising the price if inventory is low.

[1199] The server reflects the optimized pricing in the product database and transmits the new price information to the user's terminal in real time.

[1200] In addition to the demand forecasting and dynamic pricing functions provided by this system, by integrating user emotion recognition with an emotion engine, it is possible to provide a more intuitive and adaptive interface for users, which will enable companies to improve not only inventory management and production scheduling but also the user experience.

[1201] The processing flow will be explained below.

[1202] Step 1:

[1203] The server collects historical sales data, consumer behavior data, seasonality data, and market trend data from retailers and other related systems via APIs. For example, monthly sales data from the last year is retrieved from an SQL database, and customer web browsing history is collected using the Google Analytics API.

[1204] Step 2:

[1205] The server imputes missing data from the collected data, for example by imputing the average value or by using values ​​before and after the data, and then detects erroneous or outlier values ​​and corrects or deletes them as necessary.

[1206] Step 3:

[1207] The server converts numerical data to a standard scale (e.g., a range from 0 to 1) to allow comparisons between different data sets, and combines and formats data collected from different sources into a single integrated dataset.

[1208] Step 4:

[1209] The server splits the formatted dataset into training data and test data, typically using 80% of the data as training data and 20% as test data.

[1210] Step 5:

[1211] The server selects and initializes the machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) that is best suited for demand forecasting. For example, it selects the LSTM model and sets the necessary hyperparameters.

[1212] Step 6:

[1213] The server trains the model using the selected machine learning algorithm by providing training data. Specifically, it inputs past sales data into an LSTM network and builds a model to predict future sales.

[1214] Step 7:

[1215] The server inputs test data using the trained model to predict future sales trends and demand, for example, forecasting sales for the next month, and outputs the results as data.

[1216] Step 8:

[1217] The server then converts the forecast results into an appropriate format and sends them to the user's device. For example, the sales forecast results are visualized as line graphs or bar charts and displayed on the user's dashboard.

[1218] Step 9:

[1219] The device uses a camera and microphone to collect the user's facial expressions and voice data, which the emotion engine analyzes to recognize the user's emotions, such as whether the user is confused or surprised.

[1220] Step 10:

[1221] The server dynamically changes how the predictions are presented based on the recognized emotion: for example, if the user is surprised, it may display additional information to calm them down, or if the user is confused, it may display additional detailed explanations or instructions.

[1222] Step 11:

[1223] The server runs an algorithm that dynamically sets product prices based on predicted demand, for example, raising prices by 10% ahead of periods of predicted increased demand and lowering prices during periods of overstock.

[1224] Step 12:

[1225] The server takes into account the current inventory level and optimizes pricing, for example lowering the price if inventory is excessive and raising the price if inventory is low.

[1226] Step 13:

[1227] The server reflects the optimized pricing in the product database and transmits the new price information to the user's terminal in real time.

[1228] Step 14:

[1229] Users can check new price information and forecast results through their devices, and can adjust their inventory management and sales strategies accordingly.

[1230] Through this series of steps, the system of the present invention achieves highly accurate demand forecasting and dynamic pricing, and also provides an adaptive interface that responds to user emotions, allowing companies to optimize all aspects of inventory management, product production, and sales strategies.

[1231] Example 2

[1232] 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."

[1233] Conventional demand forecasting systems use past data to predict future demand, but they are unable to provide an interface that takes user emotions into account, resulting in a poor user experience. In particular, because the user's psychological state is not reflected when presenting the forecast results, there is a lack of support for the user to understand the results and take action. Conventional systems also lack the ability to dynamically set prices.

[1234] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1235] In this invention, the server includes means for collecting past data for demand forecasting, means for preprocessing the collected data, means for training a machine learning model based on the preprocessed data, means for forecasting future demand using the trained machine learning model, means for providing the forecast results to a user, means for recognizing the user's emotions, means for dynamically presenting the forecast results based on the user's emotions, and means for dynamically setting prices based on the demand forecast results. This makes it possible to provide an interface that reflects the user's emotions, improving the user experience and providing more support for understanding and taking action on the demand forecast results. Furthermore, dynamic pricing enables flexible price management in response to market fluctuations.

[1236] "Demand forecasting" is the process of predicting future consumer demand based on past and current data.

[1237] "Data Collection Measures" refers to the ability to collect necessary data from retailers and other related systems.

[1238] "Data preprocessing means" refers to functions that fill in missing values ​​in collected data, correct or delete error values ​​or outliers, and perform scaling and integration.

[1239] "Machine learning model training means" refers to a function that uses preprocessed data to train a machine learning algorithm and create a predictive model.

[1240] "Demand forecasting means" refers to the ability to forecast future demand using a trained machine learning model.

[1241] "Means of providing to users" refers to the function of providing prediction results to users in an appropriate format and presenting them in an easy-to-understand manner.

[1242] "Emotion recognition means" refers to a function that analyzes the user's facial expressions and voice to recognize the user's emotional state.

[1243] "Dynamic presentation means" refers to a function that dynamically changes the way prediction results are presented based on the recognized user emotions.

[1244] "Dynamic pricing means" refers to a function that dynamically sets product prices in real time based on demand forecast results.

[1245] This invention combines a system that uses AI and machine learning to predict future sales trends and demand with an emotion engine that recognizes user emotions. The specific processing of the program used in this system is described in detail below.

[1246] First, the server collects data. It collects past sales data, consumer behavior data, seasonal data, market trend data, etc. from retailers and related systems. Specifically, it retrieves last year's monthly sales data from an SQL database and collects users' web browsing history using the Google Analytics API. The data collected in this way includes information from a variety of sources.

[1247] Next, the server performs data preprocessing. Missing data is imputed from the collected data, and errors and outliers are detected and corrected or removed. Specific processing methods include average imputation and imputation based on surrounding values. All numerical data is also converted to a standard scale to enable comparisons between different data sets. Data collected from different sources is combined into a single integrated dataset and formatted.

[1248] The server then performs model training. The formatted dataset is divided into training data and test data, and the machine learning algorithm (e.g., LSTM, Random Forest, XGBoost, etc.) optimal for demand forecasting is selected and initialized. For example, an LSTM model is selected and the necessary hyperparameters are set. Training data is input using this selected algorithm to train the model. Specifically, past sales data is input into the LSTM network to build a model for predicting future sales.

[1249] Next, the server performs demand forecasting. It uses the trained model to input test data and predict future sales trends and demand. For example, it predicts sales for the next month and outputs the results as data.

[1250] The prediction results are provided from the server to the user's device in an appropriate format. The prediction results are visualized in line graphs and bar charts and displayed on the user's dashboard. The emotion engine uses a camera and microphone to analyze the user's facial expressions and voice when presenting the results, recognizing the user's emotions. Based on the recognized emotions, the way the prediction results are presented is dynamically changed. For example, if the user is surprised, additional information to calm them down is displayed, and if the user is confused, a detailed explanation is displayed.

[1251] Dynamic pricing is also an important feature of this system. The server runs an algorithm that dynamically sets product prices based on predicted demand. For example, the price can be raised by 10% before a period when demand is predicted to increase, and lowered when there is excess inventory. Furthermore, pricing is optimized taking into account current inventory levels. For example, if there is excess inventory, the price is lowered, and if inventory is low, the price is raised. The optimized pricing is reflected in the product database, and new price information is sent to users' devices in real time.

[1252] The system integrates demand forecasting and dynamic pricing functions, as well as user emotion recognition through an emotion engine, to provide a more intuitive and adaptive interface for users, helping companies not only to optimize inventory management and production schedules, but also to improve the user experience.

[1253] Specific examples

[1254] For example, the server retrieves last year's sales data from an SQL database and collects users' web browsing history using the Google Analytics API. It then uses mean imputation to impute missing data and scales the data to a range of 0 to 1 using MinMaxScaler. It selects an LSTM model and trains it using the training data with the fit() function. It uses the predict() function to predict next month's sales and displays the results in a line graph. The emotion engine analyzes emotions from the user's facial expressions and voice and dynamically changes the way the prediction results are presented. Based on demand, the server may raise prices by 10% or optimize prices, update the product database, and notify the user's device of the new price information.

[1255] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1256] Program processing flow and specific explanation

[1257] Step 1: Data collection

[1258] Input: Database connection information for retailers and related systems, API keys

[1259] Specific behavior:

[1260] The server uses an SQL query to retrieve last year's monthly sales data from the database.

[1261] The server uses the Google Analytics API to collect the user's web browsing history for the past six months.

[1262] Output: Collected dataset of historical sales data and consumer behavior data

[1263] Step 2: Data Preprocessing

[1264] Input: raw data collected, missing data, data containing error values ​​and outliers

[1265] Specific behavior:

[1266] The server generates a data frame using the Pandas library and performs imputation of missing values ​​(e.g., mean imputation).

[1267] The server runs an outlier detection algorithm to detect and remove or correct outliers.

[1268] The server applies MinMaxScaler to scale the numeric data to the range 0 to 1.

[1269] The server combines and formats the data collected from different sources into a single integrated data set.

[1270] Output: Preprocessed consolidated dataset

[1271] Step 3: Model training

[1272] Input: Preprocessed unified dataset, dataset split into training and test data

[1273] Specific behavior:

[1274] The server splits the combined dataset into training data (80%) and test data (20%) using the Scikit-learn or Keras libraries.

[1275] The server selects an LSTM model and sets hyperparameters (e.g., learning rate, number of epochs).

[1276] The server uses the training data to train the LSTM model using the fit() function.

[1277] Output: A trained LSTM model

[1278] Step 4: Demand forecast

[1279] Input: trained model, test data

[1280] Specific behavior:

[1281] The server uses the predict() function to forecast future sales trends and demand based on the test data.

[1282] Specifically, it generates sales forecast results for the next month.

[1283] Output: Future sales forecast results

[1284] Step 5: Output the results

[1285] Input: Future sales forecast results

[1286] Specific behavior:

[1287] The server uses the Matplotlib and Seaborn libraries to visualize the forecast results in line and bar graphs.

[1288] The server sends the visualized data to the user's device and displays it on a dashboard.

[1289] Output: Visualized data for dashboard display

[1290] Step 6: Emotion Recognition

[1291] Input: User's facial expressions, voice data

[1292] Specific behavior:

[1293] The emotion engine analyzes facial and voice data collected from the camera and microphone to recognize the user's emotional state.

[1294] The emotion engine uses facial recognition and voice analysis algorithms.

[1295] Output: Recognized emotional state of the user

[1296] Step 7: Dynamic Presentation

[1297] Input: Recognized user emotional state, future sales forecast results

[1298] Specific behavior:

[1299] The server dynamically changes the way predictions are presented based on the user's perceived emotions: if the user is surprised, it displays additional reassuring information, and if the user is confused, it displays a detailed explanation.

[1300] Output: Dynamically changed presentation results based on user emotion

[1301] Step 8: Dynamic Pricing

[1302] Input: Future sales forecast results, current inventory amount

[1303] Specific behavior:

[1304] The server dynamically sets product prices based on predicted demand, for example by raising prices by 10% before periods when demand is predicted to increase, and lowering prices when there is excess inventory.

[1305] The server optimizes pricing by taking into account inventory levels. Specifically, if inventory is excessive, the price is reduced by 5%, and if inventory is low, the price is increased by 5%.

[1306] Output: Optimized pricing information

[1307] Step 9: Update and submit pricing information

[1308] Input: Optimized Pricing Information

[1309] Specific behavior:

[1310] The server updates the product database and transmits new price information to the user terminal in real time.

[1311] Output: Updated price information, real-time notification to user terminal

[1312] (Application example 2)

[1313] 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."

[1314] Conventional demand forecasting systems primarily use historical data and machine learning to predict future sales trends, but they are unable to consider the user's emotional state. As a result, it is difficult to provide optimal interfaces and adaptive feedback for users. In particular, the lack of emotion recognition makes it difficult for users to accept and understand the prediction results, resulting in reduced usability. Furthermore, while dynamic pricing is automatically set based on demand forecasts, it fails to sufficiently increase user satisfaction because pricing based on user emotions and behavior is not taken into account.

[1315] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1316] In this invention, the server includes means for collecting past data for demand forecasting, means for preprocessing the collected data, means for training a machine learning model based on the preprocessed data, means for forecasting future demand using the trained machine learning model, means for providing the forecast results to a user, means for dynamically setting prices based on the demand forecast results, and means for recognizing the user's emotions and dynamically changing the presentation method of the forecast results and the interface based on the emotions. This makes it possible to provide more intuitive and adaptive feedback while taking the user's emotions into consideration. Dynamic pricing based on the user's emotions is also realized, thereby improving user satisfaction.

[1317] "Demand forecasting" refers to analyzing past data and predicting future sales trends and demand.

[1318] "Means for collection" refers to a function that allows necessary data to be acquired from various data sources.

[1319] "Means for data preprocessing" refers to the function of completing missing data and correcting outliers in collected data, and processing it to prepare it in a format that can be analyzed.

[1320] A "machine learning model" is a trained algorithm used to predict future demand based on historical data.

[1321] "Training" refers to the process of inputting data into a machine learning model and adjusting the model's parameters to improve its predictive accuracy.

[1322] A "demand forecasting means" is one that has the ability to forecast future sales trends and demand using trained machine learning models.

[1323] "Means for providing to the user" refers to a function that provides an interface for displaying prediction results to the user.

[1324] "Dynamic pricing" refers to the ability to adjust product prices in real time based on predicted demand.

[1325] "Means for recognizing user emotions" refers to technology that uses a camera and microphone to analyze the user's facial expressions and voice to identify their emotional state.

[1326] "Means for dynamically changing the interface presentation method" refers to a function that adaptively changes the screen display and information presentation method based on the user's emotions.

[1327] A system for implementing the present invention is configured as follows.

[1328] First, the server collects historical data for demand forecasting. The collected data includes past sales data, consumer behavior data, seasonal data, and market trend data. This data is collected by querying an SQL database or retrieving data through APIs.

[1329] The server then preprocesses the collected data, which includes imputing missing data, correcting errors and outliers, converting numerical data to a standard scale, and integrating data from different sources. This process prepares the data in a format that can be analyzed.

[1330] The server then trains a machine learning model based on the preprocessed data. The machine learning model to be used may be LSTM, Random Forest, XGBoost, or other models. The data is separated into training data and test data, and the model is trained using the selected algorithm. This process builds a highly accurate model for predicting future demand.

[1331] The server uses the trained model to predict future demand, and the predicted demand data is provided to the user. For example, the sales forecast results are displayed on the user's device as a line graph or bar chart.

[1332] The server also uses a camera and microphone to recognize the user's emotions. Using audio and image analysis software such as EmotionEngine, the server analyzes the user's facial expressions and voice to identify their emotions. Based on this identified emotion, the server dynamically changes the way prediction results are presented and the interface. For example, if the user is surprised, the server displays additional information along with the prediction result.

[1333] The server also dynamically sets prices based on the results of demand forecasts. Prices are raised if demand is predicted to increase, and lowered if there is excess inventory. This allows companies to optimize inventory management and pricing.

[1334] This system provides the following advantages: Improved usability, as users can intuitively understand the information they want; and Dynamic pricing, which maximizes profits for businesses. This is a specific embodiment of the present invention.

[1335] For example, if the user is surprised, a message will appear on the screen saying, "Don't worry. We'll show you more information." Additionally, if the current stock level is low, the product price will automatically increase.

[1336] Example prompt for generative AI model:

[1337] "How can I forecast sales trends and demand and create a dynamic interface that responds to user emotions?"

[1338] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1339] Step 1:

[1340] The server collects historical data for demand forecasting, including sales data from an SQL database, consumer behavior data, seasonality data, and market trend data through a web API. The input is data provided by various data sources, and the output is the collected raw data set.

[1341] Step 2:

[1342] The server preprocesses the collected data, which includes imputing missing data, correcting errors and outliers, and converting numerical data to a standard scale (ranging from 0 to 1). It also integrates data from different data sources. The input is the raw dataset collected in step 1, and the output is the preprocessed dataset.

[1343] Step 3:

[1344] The server trains a machine learning model based on the preprocessed data. First, it splits the data into training data and test data. Then it trains the model using a selected machine learning algorithm (e.g., LSTM, Random Forest, etc.). The input is the preprocessed dataset, and the output is the trained machine learning model.

[1345] Step 4:

[1346] The server uses the trained model to predict future demand. Test data is input to the trained model to predict future sales trends and demand. The inputs are the test data and the trained model, and the output is the predicted result data.

[1347] Step 5:

[1348] The server provides the prediction results to the user. The prediction results are visualized as line graphs or bar graphs and displayed on the user's device. The input is the prediction result data, and the output is the visual data displayed on the user's device.

[1349] Step 6:

[1350] The server uses a camera and microphone to recognize the user's emotions. It uses the EmotionEngine to analyze the user's facial expressions and voice to identify the user's emotions. The input is image and audio data obtained from the camera and microphone, and the output is the recognized emotion data.

[1351] Step 7:

[1352] The server dynamically changes the interface presentation based on the recognized emotion, for example, displaying additional information if the user is surprised, or a detailed explanation if the user is confused. The input is the recognized emotion data and the predicted result data, and the output is the adjusted interface.

[1353] Step 8:

[1354] The server dynamically sets prices based on the demand forecast results. When the forecast demand is high, the price is raised, and when there is excess inventory, the price is lowered. The input is the forecast result data and current inventory data, and the output is optimized price information.

[1355] Step 9:

[1356] The server sends new price information to the user's terminal in real time. The input is the optimized price information, and the output is the new price information displayed on the user's terminal.

[1357] 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.

[1358] 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.

[1359] 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.

[1360] 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.

[1361] 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.

[1362] 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.

[1363] 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).

[1364] 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.

[1365] 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."

[1366] 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.

[1367] 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).

[1368] 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.

[1369] 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.

[1370] 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.

[1371] 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.

[1372] 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.

[1373] 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.

[1374] 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.

[1375] 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.

[1376] 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.

[1377] 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.

[1378] The following is further disclosed regarding the above embodiment.

[1379] (Claim 1)

[1380] a means of collecting historical data for demand forecasting;

[1381] means for pre-processing the collected data;

[1382] means for training a machine learning model based on the preprocessed data;

[1383] A means of forecasting future demand using trained machine learning models; and

[1384] a means for providing the prediction result to a user;

[1385] A means for dynamically setting prices based on the demand forecast results;

[1386] A system including:

[1387] (Claim 2)

[1388] the collected data includes historical sales data, consumer behavior data, seasonality data, and market trend data;

[1389] 10. The system of claim 1.

[1390] (Claim 3)

[1391] the machine learning model uses a machine learning algorithm selected for demand forecasting;

[1392] 10. The system of claim 1.

[1393] "Example 1"

[1394] (Claim 1)

[1395] A means of collecting historical data from retailers and related systems;

[1396] A means for performing pre-processing of the collected data, such as missing data imputation, outlier detection and removal, standard scale conversion, and data combination;

[1397] A means for dividing the preprocessed data into training data and test data, and selecting and training a machine learning algorithm suitable for demand forecasting;

[1398] A means to predict future sales trends and demand using trained machine learning models;

[1399] A means for visually displaying the prediction result and transmitting it to a user's terminal;

[1400] A means for dynamically setting product prices based on demand forecast results and inventory levels, and reflecting the information in a product database;

[1401] A system including:

[1402] (Claim 2)

[1403] 10. The system of claim 1, wherein the collected data includes sales data, consumer behavior data, seasonality data, and market trend data.

[1404] (Claim 3)

[1405] 2. The system of claim 1, wherein the machine learning algorithms include LSTM, Random Forest, and XGBoost.

[1406] "Application Example 1"

[1407] (Claim 1)

[1408] a means of collecting historical data for demand forecasting;

[1409] means for pre-processing the collected data;

[1410] means for training a machine learning model based on the preprocessed data;

[1411] A means of forecasting future demand using trained machine learning models; and

[1412] a means for providing the prediction result to a user;

[1413] A means for dynamically setting prices based on the demand forecast results;

[1414] A means of forecasting sales trends and proposing appropriate price increases or reductions;

[1415] A means for visualizing the prediction results;

[1416] A system including:

[1417] (Claim 2)

[1418] the collected data includes historical sales data, consumer behavior data, seasonality data, and market trend data;

[1419] 10. The system of claim 1.

[1420] (Claim 3)

[1421] the machine learning model uses a machine learning algorithm selected for demand forecasting;

[1422] 10. The system of claim 1.

[1423] "Example 2: Combining Emotion Engines"

[1424] (Claim 1)

[1425] a means of collecting historical data for demand forecasting;

[1426] means for pre-processing the collected data;

[1427] means for training a machine learning model based on the preprocessed data;

[1428] A means of forecasting future demand using trained machine learning models; and

[1429] a means for providing the prediction result to a user;

[1430] means for recognizing a user's emotion;

[1431] A means for dynamically presenting a prediction result based on a user's emotion;

[1432] A means for dynamically setting prices based on the demand forecast results;

[1433] A system including:

[1434] (Claim 2)

[1435] the collected data includes historical sales data, consumer behavior data, seasonality data, and market trend data;

[1436] 10. The system of claim 1.

[1437] (Claim 3)

[1438] the machine learning model uses a machine learning algorithm selected for demand forecasting;

[1439] 10. The system of claim 1.

[1440] "Application example 2 when combining emotion engines"

[1441] (Claim 1)

[1442] a means of collecting historical data for demand forecasting;

[1443] means for pre-processing the collected data;

[1444] means for training a machine learning model based on the preprocessed data;

[1445] A means of forecasting future demand using trained machine learning models; and

[1446] a means for providing the prediction result to a user;

[1447] A means for dynamically setting prices based on the demand forecast results;

[1448] means for recognizing a user's emotion and dynamically changing the prediction results and the interface presentation method based on the emotion;

[1449] A system including:

[1450] (Claim 2)

[1451] the collected data includes historical sales data, consumer behavior data, seasonality data, and market trend data;

[1452] 10. The system of claim 1.

[1453] (Claim 3)

[1454] the machine learning model uses a machine learning algorithm selected for demand forecasting;

[1455] 10. The system of claim 1. [Explanation of symbols]

[1456] 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 of collecting historical data for demand forecasting; means for pre-processing the collected data; means for training a machine learning model based on the preprocessed data; A means of forecasting future demand using trained machine learning models; and a means for providing the prediction result to a user; A means for dynamically setting prices based on the demand forecast results; A system including:

2. the collected data includes historical sales data, consumer behavior data, seasonality data, and market trend data; The system of claim 1 .

3. the machine learning model uses a machine learning algorithm selected for demand forecasting; The system of claim 1 .

Images