System

Abstract

A system is provided.SOLUTION: A system, comprising: means for collecting historical sales data, weather data, and economic indicator data; means for pre-processing the collected data, supplementing missing values, and normalizing; means for training a machine learning model based on the pre-processed data; means for inputting the latest data and generating a sales forecast in response to a sales forecast request from a user; and means for displaying the generated sales forecast result on the user's terminal and visually showing it on a dashboard.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 many companies, sales management and forecasting rely on manual processes, resulting in unrealistic sales targets and excess inventory, which increases the workload of sales representatives. As a result, the pressure of not achieving sales targets increases, making efficient sales activities difficult. Furthermore, excess inventory has a negative impact on the environment and wastes resources. The purpose of this invention is to solve these problems and achieve efficient sales activities and environmental protection. [Means for solving the problem]

[0005] The present invention is a system that includes a means for collecting past sales data, weather data, and economic indicator data, a means for preprocessing the collected data, imputing missing values, and normalizing the data, a means for training a machine learning model based on the preprocessed data, a means for inputting the latest data and generating a sales forecast in response to a sales forecast request from a user, and a means for displaying the generated sales forecast results on the user's terminal and visually showing them on a dashboard. The system also includes a means for simultaneously proposing inventory management based on past sales data, the latest weather data, and economic trend data, and a means for making the sales forecast results available for download as a PDF report. This improves the efficiency of sales activities, optimizes inventory management, and contributes to environmental protection.

[0006] "Sales data" refers to records of product sales within a certain period of time at a company, and specifically includes information such as the sales date, product ID, sales quantity, price, and area.

[0007] "Weather data" refers to information about past and present weather conditions provided by meteorological agencies, including, specifically, temperature, precipitation, wind speed, and air pressure.

[0008] "Economic indicator data" refers to numerical indicators that show the state of the economy, and includes information such as GDP growth rate, unemployment rate, and consumer confidence index.

[0009] "Means of collection" refers to the methods and technologies used to obtain data from external or internal sources, including, for example, acquisition using APIs and extraction from databases.

[0010] "Preprocessing means" refers to operations used to prepare collected data in a format suitable for machine learning models, and specifically includes techniques such as data normalization, missing value imputation, and scaling.

[0011] A "machine learning model" is an algorithm that learns patterns from data and makes predictions and classifications based on the learning results. Examples include random forest regression and LSTM (long short-term memory) networks.

[0012] "User's terminal" refers to the computing device used by the user to access the sales forecasting system, and specifically includes a PC, tablet, smartphone, etc.

[0013] "Sales forecast results" are results of forecasting future sales using a machine learning model, and include predicted sales, peak demand times, and details by region.

[0014] A "dashboard" refers to an interface that visually displays information in a user-accessible manner, for example, by showing data in the form of graphs, charts, tables, etc.

[0015] "Inventory management proposal" refers to instructions and guidelines for proposing appropriate inventory levels and purchasing timing based on forecasted sales data and actual inventory status.

[0016] "PDF format" is an abbreviation for Portable Document Format, and is a file format widely used for electronic documents. [Brief explanation of the drawings]

[0017] [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

[0018] 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.

[0019] First, the terms used in the following description will be explained.

[0020] 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).

[0021] 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.

[0022] 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.

[0023] 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.

[0024] 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."

[0025] [First embodiment]

[0026] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0027] 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.

[0028] 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).

[0029] 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.

[0030] 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.

[0031] 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.

[0032] 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.

[0033] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0034] 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.

[0035] 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.

[0036] 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.

[0037] 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."

[0038] The present invention is a system that collects past sales data, weather data, and economic indicator data, and uses machine learning models to forecast sales based on this data. This system is designed to improve sales efficiency and protect the environment, and an embodiment of the system is described below.

[0039] System Overview

[0040] This system consists of a server, a user's device, and a network infrastructure that connects them. The server has multiple functions, such as data collection, preprocessing, model training, prediction generation, and result display. Meanwhile, users input requests through their device and check the prediction results.

[0041] Program processing flow

[0042] 1. Data Collection

[0043] The server periodically collects sales data, weather data, and economic indicator data. The sales data is obtained from an internal database, the weather data is obtained from a weather agency's API, and the economic indicator data is obtained from an economic database.

[0044] 2. Data Preprocessing

[0045] The server preprocesses the collected data by imputing missing values, removing noise, scaling, and normalizing it. For example, for periods where sales data is missing, it uses similar data from the past to impute the missing data.

[0046] 3. Model Training

[0047] The server uses the preprocessed data to train a machine learning model, which uses a selected algorithm (e.g., random forest regression or LSTM) for sales forecasting. The model learns sales patterns based on past data.

[0048] 4. Prediction Generation

[0049] When a user sends a sales forecast request from their device, the server retrieves the latest weather data and economic trend data from the API and uses this data to generate a sales forecast in real time.

[0050] 5. Display results

[0051] The server generates the prediction results and sends them to the user's device, where they can view the results visually on a dashboard and optionally download the report in PDF format.

[0052] Specific examples

[0053] Data collection

[0054] The server retrieves sales data for the past five years from an internal database, weather data (temperature, precipitation, etc.) for the past five years from the Japan Meteorological Agency via an API, and economic indicator data (e.g., GDP growth rate, unemployment rate) from the government and reliable economic forums.

[0055] Data Preprocessing

[0056] The server fills in missing data from the collected sales data using past approximate data. It also scales the sales data and weather data to a range of 0 to 1 and converts them into a format suitable for the model.

[0057] Model learning

[0058] The server trains a random forest regression model on the preprocessed data, which is used to learn about past sales patterns and the effects of weather and economic conditions and to predict future sales.

[0059] Forecast Generation

[0060] When a user requests "Sales forecast for October 2023," the server retrieves the latest weather forecast and economic indicators and inputs them into the trained model, which then outputs detailed sales forecasts and supply and demand plans for each region.

[0061] Results display

[0062] The server organizes the generated sales forecast results and displays them on the user's dashboard, where users can view the data in graphs and tables and download the reports in PDF format if required.

[0063] This frees sales representatives from manual work, enabling more efficient sales activities and achieving sustainable inventory management across the entire company.

[0064] The processing flow will be explained below.

[0065] Step 1:

[0066] The server retrieves past sales data from an internal database, which stores information such as sales date, product ID, sales quantity, price, and area.

[0067] Step 2:

[0068] The server obtains weather data from the weather agency's API, including past temperature, precipitation, wind speed, and air pressure.

[0069] Step 3:

[0070] The server retrieves economic indicator data from economic databases or trusted government APIs, such as GDP growth, unemployment rate, and consumer confidence index.

[0071] Step 4:

[0072] The server aggregates the collected sales, weather, and economic data, making the data set available in a consistent manner.

[0073] Step 5:

[0074] The server preprocesses the integrated data, imputing missing values, removing noise, and scaling the data. For example, if there is missing sales data, it is imputed using similar data from the past.

[0075] Step 6:

[0076] The server normalizes the data by scaling each variable to a range between 0 and 1. This step is important for improving the accuracy of training machine learning models.

[0077] Step 7:

[0078] The server uses the preprocessed data to train machine learning models, for example, using random forest regression or LSTM (long short-term memory) networks.

[0079] Step 8:

[0080] The server uses validation data to optimize model parameters during training and prevent overfitting or underfitting, a process that evaluates how well the model performs on real-world data.

[0081] Step 9:

[0082] The server compares the predictions with actual sales data to evaluate the accuracy of the model and tunes the model as needed, a step that is important for continuously improving the model's accuracy.

[0083] Step 10:

[0084] A user enters a request into a terminal, such as "Sales forecast for October 2023," which starts the forecast generation process.

[0085] Step 11:

[0086] The server retrieves the latest weather and economic indicator data in real time and inputs this into the trained model.

[0087] Step 12:

[0088] The server generates a sales forecast, which includes projected sales figures, peak demand times, and geographical details.

[0089] Step 13:

[0090] The server also generates inventory management recommendations based on the generated forecasts, helping to prevent overstocks and out-of-stock situations.

[0091] Step 14:

[0092] The server sends the prediction results and inventory management suggestions to the user's terminal.

[0093] Step 15:

[0094] The device visually displays the received prediction results on a dashboard, allowing users to view the results in graph and table format.

[0095] Step 16:

[0096] Users can check the forecast results on the dashboard and download the report in PDF format as needed, which allows them to develop specific sales strategies and inventory management plans.

[0097] As described above, the system collects data, preprocesses it, trains models, generates predictions, and displays the results step by step, supporting efficient sales activities and sustainable inventory management.

[0098] Example 1

[0099] 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."

[0100] Conventional sales forecasting systems often make predictions based solely on past sales data, without taking into account multiple factors such as weather data and economic indicators, resulting in low accuracy. Another issue is that few systems provide sales forecast results in real time, and the effects of missing data and noise cannot be fully mitigated. Furthermore, there is a lack of functionality that allows users to easily check sales forecast results and provides them in a format that can be used as reference material. These issues make it difficult to improve sales efficiency and protect the environment.

[0101] 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.

[0102] In this invention, the server includes: means for periodically collecting past sales data, weather data, and economic indicator data from an internal company database or an external API; means for preprocessing the collected data, filling in missing values ​​with past approximate data, removing noise, scaling, and normalizing the data; means for training a machine learning model based on the preprocessed data to learn past sales patterns and environmental factors; means for inputting the latest weather data and economic indicator data in response to a sales forecast request from a user and generating a sales forecast in real time; and means for transmitting the generated sales forecast results to the user's device, visually displaying them on a dashboard, and making them available for download as a PDF report if necessary. This improves the accuracy of sales forecasts and enables real-time information provision and visual display of results. As a result, sales efficiency and sustainable inventory management can be achieved.

[0103] An "in-house database" is a system installed within a company that stores and manages information, and holds business data such as sales data and customer information.

[0104] An "external organization's API" is an API (Application Programming Interface) provided by an external organization or service, and is a means of programmatically obtaining information such as weather data and economic indicator data.

[0105] "Sales data" refers to numerical information about sales recorded by a company during a specific period, such as the sales quantity and sales amount of a product.

[0106] "Weather data" is numerical information about the weather in a specific region or period, such as temperature, precipitation, humidity, and wind speed.

[0107] "Economic indicator data" refers to various statistical data that show the state of the economy, such as GDP growth rate, unemployment rate, and consumer price index.

[0108] "Preprocessing" refers to the preparation of data prior to data analysis or machine learning, and specifically includes processes such as filling in missing values, removing noise, scaling, and normalization.

[0109] "Missing values" refers to a state in which some data is missing, including blanks and NULL values ​​in a dataset.

[0110] "Noise" refers to unnecessary disturbances and errors contained in data that can reduce the accuracy of analyses and models.

[0111] "Scaling" is a conversion technique used to fit data values ​​into a range, usually normalizing the data range to 0 to 1 using minimum and maximum values.

[0112] "Normalization" refers to a transformation process to make the distribution of data uniform, particularly by confining the data to a specific range through scaling.

[0113] A "machine learning model" is a set of algorithms that learn patterns and rules from data and make predictions or classifications. Examples include random forests and LSTM.

[0114] "Training" is the process by which a machine learning model uses data to learn optimal parameters.

[0115] A "sales forecast request" is a request sent by a user to a server to forecast future sales based on a specific period and conditions.

[0116] "Real-time" refers to a state in which data is processed and displayed immediately at the moment it is generated or acquired.

[0117] A "dashboard" is a visual interface that allows users to see important information at a glance, displaying data in the form of graphs and tables.

[0118] "PDF format" stands for Portable Document Format, a file format for storing and sharing information in a fixed-layout format.

[0119] A "report" refers to the analysis results or report on specific data, and is often generated in PDF format or similar.

[0120] This invention is a system that collects past sales data, weather data, and economic indicator data from in-house databases and APIs of external institutions, and uses this data to perform sales forecasts using machine learning models. This system consists of a server, user terminals, and a network infrastructure that connects these.

[0121] Hardware and Software Configuration

[0122] server

[0123] The server is a computer system equipped with a high-performance processor and large memory capacity, and is installed with software for database management systems, API handlers, and machine learning models, including pandas, scikit-learn, and TensorFlow for data analysis.

[0124] Terminal

[0125] Users access the system through a web browser on devices such as PCs and tablets. The device must have an internet connection and a browser (e.g., Google Chrome or Mozilla Firefox) to view dashboards and reports.

[0126] Data collection

[0127] The server periodically connects to the company's internal database and retrieves sales data for the past five years, which is retrieved using SQL queries.

[0128] The server retrieves weather data for the past five years through the API of an external weather agency. This API request uses the HTTP GET method to retrieve data such as temperature and precipitation for a specified period.

[0129] The server uses APIs provided by governments and economic forums to collect economic indicator data (e.g., GDP growth rate, unemployment rate) for the past five years.

[0130] Data Preprocessing

[0131] The server uses the pandas library to fill in missing values ​​in the acquired sales data with similar historical data, and applies appropriate filtering techniques to remove noise.

[0132] Next, we use scikit-learn's MinMaxScaler to scale and normalize the sales and weather data to the range 0 to 1, so that the data can be fed into the machine learning model in a unified format.

[0133] Model learning

[0134] The server trains a machine learning model based on the preprocessed data. Algorithms such as random forest regression models and LSTM are used, and the model is run using scikit-learn and TensorFlow. During the training process, the dataset is split into training data and validation data, and cross-validation is applied to ensure the accuracy of the model.

[0135] Forecast Generation

[0136] When a user submits a sales forecast request from their device, they enter a prompt, such as "Sales forecast for October 2023." In response, the server again retrieves the latest weather and economic data and inputs it into the trained model.

[0137] The model generates results by predicting future sales in real time based on new data.

[0138] Results display

[0139] The server organizes the generated prediction results and sends them to the user's device, where they are visually displayed in graphs and tables on a dashboard for the user to review.

[0140] If desired, users can download the report in PDF format, a feature that uses libraries such as matplotlib and ReportLab.

[0141] Specific examples

[0142] Data collection

[0143] The server retrieves sales data for the past five years from a company database, collects weather data from the Japan Meteorological Agency's API, and obtains economic indicator data from the government's economic forum.

[0144] Data Preprocessing

[0145] The server fills in missing sales data with similar data from the past and scales the weather data to normalize it to the range 0 to 1.

[0146] Model learning

[0147] The server trains a random forest regression model to learn patterns from the data.

[0148] Forecast Generation

[0149] When a user requests "Sales forecast for October 2023," the server inputs the latest data into the model and generates a sales forecast.

[0150] Results display

[0151] The server displays the prediction results on a dashboard and allows users to download the report in PDF format.

[0152] Prompt Sentence Examples

[0153] "Show me sales forecast for October 2023."

[0154] "Get weather data for the past five years."

[0155] "Please update your economic indicator data."

[0156] This system frees sales representatives from manual work, enabling more efficient sales activities and achieving sustainable inventory management across the entire company.

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

[0158] Step 1:

[0159] The server connects to the company's internal database and retrieves sales data for the past five years. This is done using an SQL query. Specifically, the server executes a query such as "SELECT FROM sales WHERE date BETWEEN '2018-01-01' AND '2022-12-31'" to extract the required data. The input is the connection information and the sales data contained in the database, and the output is the sales data for the specified period.

[0160] Step 2:

[0161] The server sends an HTTP GET request to the weather agency's API to retrieve weather data for the past five years. Specifically, it sends a request such as "GET http: / / api.weather.com / data?startDate=2018-01-01&endDate=2022-12-31" to retrieve temperature and precipitation data. The input is the API endpoint information, and the output is the weather data returned by the API.

[0162] Step 3:

[0163] The server accesses the government's economic data API to obtain economic indicator data (e.g., GDP growth rate and unemployment rate) for the past five years. Specifically, it executes "GET http: / / api.economicdata.com / indicators?startDate=2018-01-01&endDate=2022-12-31" to obtain the indicator data. The input is the API endpoint information, and the output is the obtained economic indicator data.

[0164] Step 4:

[0165] The server preprocesses the collected sales data, weather data, and economic indicator data. Specifically, it creates a data frame using the pandas library and imputes missing values ​​using methods such as linear imputation. The input is the collected raw data, and the output is a data frame with missing values ​​imputed.

[0166] Step 5:

[0167] The server scales and normalizes the preprocessed data, specifically using scikit-learn's MinMaxScaler to convert values ​​to the range 0 to 1. The input is the interpolated data frame, and the output is the scaled data.

[0168] Step 6:

[0169] The server uses the scaled data to train a machine learning model, for example, a random forest regression model using scikit-learn's RandomForestRegressor. The input is the preprocessed and scaled data, and the output is a trained machine learning model.

[0170] Step 7:

[0171] The user sends a sales forecast request from the terminal. For example, they input a prompt such as "Sales forecast for October 2023." The input is the user's request, and the output is the request data to the server.

[0172] Step 8:

[0173] The server retrieves the latest weather data and economic indicator data from the API again and inputs it into the trained model, generating a sales forecast in real time. The input is the latest weather data and economic indicator data, and the output is sales forecast data.

[0174] Step 9:

[0175] The server organizes the generated sales forecast results and sends them to the user's device. The data is displayed visually on a dashboard. The input is the sales forecast data, and the output is the graphs and tables displayed on the dashboard.

[0176] Step 10:

[0177] Users can check the forecast results through a dashboard and download them in PDF format if necessary. The server generates reports using libraries such as matplotlib and ReportLab. The input is the user's download request and sales forecast data, and the output is a PDF report.

[0178] (Application example 1)

[0179] 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."

[0180] In traditional manufacturing, demand forecasting and inventory management are often performed manually, resulting in problems such as human error and reduced efficiency. It is also difficult to flexibly adjust manufacturing processes based on real-time data from the field, making it difficult to maintain a balance between supply and demand. Furthermore, it is difficult to make accurate predictions that take environmental fluctuations (weather and economic conditions) into account, leading to inventory surpluses and shortages. This calls for a method to improve production efficiency, minimize resource waste, and establish a sustainable production system.

[0181] 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.

[0182] In this invention, the server includes: means for collecting past sales data, weather data, and economic index data; means for preprocessing the collected data, imputing missing values, and normalizing the data; means for training a machine learning model based on the preprocessed data; means for inputting the latest data and generating a sales forecast in response to a sales forecast request from a user; means for displaying the generated sales forecast results on the user's terminal and visually showing them on a dashboard; and means for instructing a robot to adjust the manufacturing process based on the generated forecast results. This allows the factory's manufacturing process to be automatically adjusted in real time based on supply and demand forecasts, not only improving production efficiency but also reducing human error and enabling sustainable inventory management.

[0183] "Past sales data" refers to sales performance information recorded by companies and stores to date, specifically information such as product sales quantity, sales amount, and sales period.

[0184] "Weather data" refers to meteorological information such as temperature, precipitation, and wind speed obtained from meteorological agencies and used in forecasting models.

[0185] "Economic indicator data" refers to data that shows economic trends, and includes GDP growth rate, unemployment rate, consumer price index, etc.

[0186] "Means of collection" refers to the methods and technologies used to obtain the necessary information from various data sources.

[0187] "Preprocessing means" refers to methods and techniques for converting acquired data into an analyzable format, and specifically includes imputing missing values ​​and normalizing data.

[0188] "Methods for imputing missing values" are methods for completing incomplete data by using similar data from the past or average values, etc.

[0189] "Normalization" is a method of scaling data values ​​into a certain range so that machine learning algorithms can learn efficiently.

[0190] A "means for training a machine learning model" is a method for training a machine learning algorithm using preprocessed data to build a predictive model.

[0191] The "means for generating sales forecasts" is a method for inputting the latest data into a trained machine learning model to predict future sales.

[0192] A "user's terminal" is a device such as a PC or smartphone that the user uses to check the prediction results.

[0193] A "dashboard" refers to an interface that allows users to visually check prediction results, displaying information in the form of graphs and tables.

[0194] "Means for issuing instructions to robots to adjust the manufacturing process based on the generated prediction results" refers to methods and technologies for issuing specific operating instructions to manufacturing robots based on predicted supply and demand data, automatically adjusting the manufacturing line.

[0195] The present invention is a system that collects past sales data, weather data, and economic indicator data, and uses a machine learning model based on this data to forecast sales, and this system has been applied to a factory robot. This system is configured as follows.

[0196] System Overview

[0197] This system consists of a server, user terminals, factory robots, and a network infrastructure that connects them. The server has multiple functions, including data collection, preprocessing, model learning, prediction generation, and result display. Meanwhile, users input requests through terminals or consoles and check prediction results. Furthermore, the factory robots automatically adjust the manufacturing process based on this prediction data.

[0198] Hardware and software used

[0199] Hardware:

[0200] Factory robots (e.g. Fanuc, ABB, KUKA, etc.)

[0201] Servers (cloud or on-site data center)

[0202] Sensors (monitor temperature, humidity, vibration, etc.)

[0203] software:

[0204] Machine learning frameworks (e.g. TensorFlow, PyTorch)

[0205] Database (e.g. MySQL, PostgreSQL)

[0206] API services (e.g. OpenWeatherMap, economic data API)

[0207] Robot control software (e.g. ROS - Robot Operating System)

[0208] Data processing and calculation

[0209] 1. The server retrieves past sales data from the company's internal database, collects weather data from the weather agency's API, and collects economic indicator data from the economic database.

[0210] 2. The server preprocesses the collected data. Specifically, it imputes missing values, removes noise, and scales and normalizes the data. For example, if sales data is missing for a period, it is imputed using similar data from the past.

[0211] 3. The server uses the preprocessed data to train a machine learning model. This model uses a selected algorithm (e.g., random forest regression or LSTM) for sales forecasting. The model learns sales patterns based on past data.

[0212] 4. When a user sends a sales forecast request from their device, the server retrieves the latest weather data and economic trend data from the API and uses this data to generate a sales forecast in real time.

[0213] 5. The server generates the prediction results and sends them to the user's device. The user can view the results visually on a dashboard and optionally download them as a PDF report.

[0214] 6. Based on the generated prediction results, the server sends instructions to the factory robot to adjust the manufacturing process, allowing the factory robot to automatically optimize the manufacturing process based on the supply and demand balance.

[0215] Specific examples

[0216] Data collection:

[0217] The server retrieves the past two years of production and inventory data from the factory database, and also uses APIs to retrieve the past two years of weather data (e.g., temperature, humidity) and economic data (e.g., GDP growth rate, production index).

[0218] Data preprocessing:

[0219] The server fills in missing data using past approximate data, scales sales data and weather data to a range of 0 to 1, and converts them into a format suitable for the model.

[0220] Model training:

[0221] The server trains a random forest regression model on the preprocessed data, which is used to learn about past sales patterns and the effects of weather and economic conditions and to predict future sales.

[0222] Forecast generation:

[0223] When a user requests "Sales forecast for October 2023," the server retrieves the latest weather forecast and economic indicators and inputs them into the trained model, which then outputs detailed sales forecasts and supply and demand plans for each region.

[0224] Results display and automatic adjustment:

[0225] The server then organizes the generated sales forecasts and displays them on the user's dashboard. Based on the forecasts, the server sends instructions to factory robots to adjust the manufacturing process, for example, adjusting the production speed of a specific product line to prevent overstocks or shortages.

[0226] Example prompts to input to the generative AI model

[0227] Input historical production data, weather data, and economic indicator data to generate a production demand forecast for the next month.

[0228] Using this system, supply and demand forecasts and production process optimization at manufacturing sites can be performed in real time, improving production efficiency and optimizing inventory management.

[0229] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0230] Step 1: Data collection

[0231] The server retrieves past sales data from the company's internal database, and uses APIs to collect weather data from meteorological agencies and economic indicator data from economic databases. As a result, past sales data, weather data, and economic indicator data are aggregated on the server as input data. Specifically, this includes sales quantity, sales amount, temperature, precipitation, GDP growth rate, unemployment rate, etc.

[0232] Step 2: Data Preprocessing

[0233] The server then fills in missing values ​​in the collected data, removes noise, and scales and normalizes the data. Specifically, it fills in missing sales data using similar historical data, and does the same for weather data and economic indicator data. The result is a preprocessed dataset that can be used by machine learning models.

[0234] Step 3: Train the machine learning model

[0235] The server uses the preprocessed data to train a machine learning model for sales forecasting. Specifically, it uses algorithms such as random forest regression and LSTM to learn sales patterns and correlations with various metrics, resulting in a trained machine learning model.

[0236] Step 4: Receive a prediction request

[0237] The user sends a sales forecast request from the terminal. For example, they enter a prompt sentence such as "Please generate sales forecast for October 2023." This becomes the input data to the server.

[0238] Step 5: Generate forecasts

[0239] The server receives the request from the user and retrieves the latest weather and economic indicator data from the API again. The retrieved data is input into the trained model to generate a sales forecast. This operation outputs real-time sales forecast data.

[0240] Step 6: View the results

[0241] The server sends the generated sales forecast results to the user's device, where the user can visually check the forecast results on the dashboard of their device. For example, the sales forecast data can be displayed in graph or table format.

[0242] Step 7: Adjust the manufacturing process based on the results

[0243] Based on the generated predictions, the server sends instructions to the factory robots to adjust the manufacturing process. The factory robots then use the instructions received from the server to optimize the manufacturing process in real time, for example, by adjusting the production speed of a particular product line or changing the amount of materials required.

[0244] Through these processing steps, the system automates everything from sales forecasts to real-time adjustments to manufacturing processes, enabling efficient production management.

[0245] 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.

[0246] This invention combines a system that collects past sales data, weather data, and economic indicator data and uses machine learning models to forecast sales, with an emotion engine that recognizes user emotions. This system is designed to improve sales efficiency and protect the environment, and also enables flexible responses based on user emotions.

[0247] System Overview

[0248] This system consists of a server, a user's device, an emotion engine, and a network infrastructure that connects these components. The server has multiple functions, including data collection, preprocessing, model training, prediction generation, result display, and emotion recognition. Meanwhile, users input requests through their device and check prediction results.

[0249] Program processing flow

[0250] 1. Data Collection

[0251] The server periodically collects historical sales data, weather data, and economic indicator data. The sales data is obtained from an internal database, the weather data is obtained from a weather agency's API, and the economic indicator data is obtained from an economic database.

[0252] 2. Data Preprocessing

[0253] The server preprocesses the collected data by filling in missing values, removing noise, scaling, and normalizing it. For example, if there are gaps in sales data, it fills them in using similar data from the past.

[0254] 3. Model Training

[0255] The server uses the preprocessed data to train a machine learning model using techniques such as random forest regression and LSTM (long short-term memory) networks. The model learns sales patterns based on past data.

[0256] 4. Emotion recognition

[0257] The device analyzes the user's facial expressions and tone of voice, and uses an emotion engine to recognize the user's emotions. The emotion data is sent to the server in real time.

[0258] 5. Prediction Generation

[0259] When a user sends a sales forecast request from their device, the server retrieves the latest weather and economic indicator data, inputs this data into the trained model, and adjusts the forecast taking into account emotion data from the emotion engine.

[0260] 6. Results display

[0261] The server generates prediction results and inventory management suggestions and sends them to the user's device. The user can view the results visually on a dashboard, and emotion data is used to dynamically adjust the display content and interface.

[0262] Specific examples

[0263] Data collection

[0264] The server retrieves sales data for the past five years from an internal database, and also retrieves weather data (temperature, precipitation, etc.) for the past five years from meteorological agencies via API, as well as economic indicator data (GDP growth rate, unemployment rate).

[0265] Data Preprocessing

[0266] The server normalizes the collected data by imputing missing values, removing noise, and scaling. Missing sales data is interpolated based on similar patterns in the past.

[0267] Model learning

[0268] The server trains a random forest regression model on the pre-processed data to learn about past sales patterns and the effects of weather and economic conditions.

[0269] emotion recognition

[0270] The device analyzes the user's facial expressions and voice tone in real time and obtains emotional data through an emotion engine, which determines whether the user is, for example, stressed or relaxed.

[0271] Forecast Generation

[0272] When a user requests "Sales forecast for October 2023," the server retrieves the latest weather forecast, economic indicator data, and sentiment data, and inputs these into the model to generate a sales forecast. If the server determines that the user is stressed, it will carefully provide feedback on the forecast and present the results.

[0273] Results display

[0274] The server sends the forecast results and inventory management suggestions to the user's device, which displays them in graph and table format on a dashboard. The user can check the results and download the report in PDF format if necessary. The display content is also adjusted and the interface is customized based on the user's sentiment.

[0275] In this way, the present invention can provide more accurate sales forecasts and inventory management while responding to user emotions.

[0276] The processing flow will be explained below.

[0277] Step 1:

[0278] The server retrieves past sales data from an internal database, which stores information such as sales date, product ID, sales quantity, price, and area.

[0279] Step 2:

[0280] The server obtains weather data from the weather agency's API, specifically, past temperature, precipitation, wind speed, air pressure, etc.

[0281] Step 3:

[0282] The server retrieves economic indicator data from economic databases or APIs of trusted government agencies, such as GDP growth, unemployment rate, and consumer confidence index.

[0283] Step 4:

[0284] The server aggregates the collected sales data, weather data, and economic indicator data, allowing for a consistent analysis of the data set.

[0285] Step 5:

[0286] The server preprocesses the integrated data by imputing missing values, removing noise, and scaling it. For example, if there is missing sales data, it fills in the missing data using similar data from the past.

[0287] Step 6:

[0288] The server uses the preprocessed data as training data for machine learning models, such as random forest regression and LSTM (long short-term memory) networks, to train the models.

[0289] Step 7:

[0290] The server uses validation data to optimize the parameters of the trained model and prevent overfitting and underfitting.

[0291] Step 8:

[0292] The device analyzes the user's facial expressions and voice tone in real time and recognizes their emotions using an emotion engine. The recognized emotion data is then sent to the server.

[0293] Step 9:

[0294] When a user requests "Sales forecast for October 2023" from their device, the server retrieves the latest weather data, economic indicator data, and sentiment data.

[0295] Step 10:

[0296] The server then inputs this data into a trained model to generate sales forecasts, with the sentiment data being used to fine-tune the results.

[0297] Step 11:

[0298] The server organizes the generated sales forecast results and simultaneously creates inventory management proposals, thereby preventing excess inventory and out-of-stock situations.

[0299] Step 12:

[0300] The server transmits the sales forecast results and inventory management proposals to the user's terminal.

[0301] Step 13:

[0302] The terminal displays the received sales forecast results and inventory management suggestions on a dashboard, including graphs and tables.

[0303] Step 14:

[0304] Users can check the forecast results on the dashboard and download the report in PDF format if necessary.

[0305] Step 15:

[0306] The device dynamically adjusts the display content and interface based on the user's emotions. For example, if the user is feeling stressed, the device will change the display to a simpler display and a more visually soothing color scheme.

[0307] As described above, the system performs data collection, preprocessing, model training, emotion recognition, prediction generation, and result display at each step, supporting efficient sales activities and sustainable inventory management while also improving the user experience.

[0308] Example 2

[0309] 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."

[0310] Conventional sales forecasting systems typically collect past sales data, weather data, economic indicator data, etc., and use machine learning to forecast sales. However, they were unable to take into account individual factors such as the user's emotional state. This meant that they were unable to respond flexibly to user emotions, making it difficult to improve forecast accuracy and user experience. Furthermore, there were limited ways to visually confirm forecast results and present them in a format that was easy for users to understand.

[0311] 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.

[0312] In this invention, the server includes: means for collecting past sales data, weather data, and economic indicator data; means for preprocessing the collected data, imputing missing values, and normalizing the data; means for training a machine learning model based on the preprocessed data; means for recognizing user emotion data and transmitting it to the server; means for inputting the latest data and emotion data and generating a sales forecast in response to a sales forecast request from the user; and means for displaying the generated sales forecast results on the user's terminal and visually presenting them on a dashboard. This enables sales forecasts that take the user's emotional state into account, improving forecast accuracy and the user experience. Furthermore, visually displaying the forecast results makes it possible to provide a system that is easy for users to understand and use.

[0313] "Past sales data" refers to records of past sales and purchases of goods conducted by a company, and includes data such as sales quantities, amounts, and dates.

[0314] "Weather data" refers to data including meteorological information such as temperature, precipitation, wind speed, and humidity obtained from meteorological agencies.

[0315] "Economic indicator data" is statistical data used to understand the state of the economy, such as gross domestic product (GDP), unemployment rate, and inflation rate.

[0316] "Preprocessing" refers to the processing performed on collected raw data, and involves cleansing and normalizing the data, such as filling in missing values, removing noise, and scaling.

[0317] A "machine learning model" is an algorithm or computational model that learns patterns and regularities from given data and makes predictions and judgments about future data.

[0318] "Emotion data" is data that indicates the emotional state of the user analyzed from facial expressions, tone of voice, and the like.

[0319] A "server" is a central computer that performs data collection, preprocessing, model training, prediction generation, and display of results.

[0320] A "terminal" is a computer or mobile device used by a user to enter data and confirm results.

[0321] A "dashboard" is an interface that allows users to visually check information such as sales forecast results and inventory management proposals.

[0322] This invention is a system that collects and analyzes past sales data, weather data, economic indicator data, and even user emotion data, and uses machine learning models to forecast sales. This system is composed of a server, a terminal, an emotion recognition engine, and a network infrastructure that links these components. Below, we will explain in detail how each component works.

[0323] Data collection

[0324] The server periodically collects past sales data, weather data, and economic indicator data. Sales data is obtained from an internal database, weather data is obtained using the weather agency's API, and economic indicator data is obtained from an economic database. Specifically, sales data is extracted from the database by executing SQL queries using Python. Weather data is obtained in JSON format from the weather agency's API using an HTTP request, for example, and economic indicator data is collected from the IMF and OECD APIs.

[0325] Data Preprocessing

[0326] The server preprocesses the collected data. This preprocessing includes imputing missing values, removing noise, scaling, and normalizing the data. For example, it uses Python's pandas library to impute missing values ​​in the data frame with the mean value and scales the data using scikit-learn's StandardScaler. This cleans the data so that the machine learning model can operate optimally.

[0327] Model learning

[0328] The server uses the preprocessed data to train a machine learning model. The main techniques used are random forest regression and LSTM (long short-term memory). For example, a model is created using scikit-learn's RandomForestRegressor and trained using the training data. The server also evaluates the accuracy of the model and adjusts hyperparameters using cross-validation if necessary.

[0329] emotion recognition

[0330] The device captures the user's facial expressions and voice tone and analyzes them with an emotion recognition engine. For example, it uses OpenCV to analyze video from a webcam and recognize emotions from facial expressions. It also uses the Librosa library to extract voice features and classify emotions using a deep learning model. This emotion data is sent to a server in real time and incorporated into a prediction algorithm.

[0331] Forecast Generation

[0332] When a user requests a sales forecast through their device, the server retrieves the latest weather data, economic indicator data, and emotion data and inputs this into the trained model. For example, if a user requests "Sales forecast for October 2023," the server collects the data and generates a sales forecast. The user's emotion data is also taken into account and the forecast results are adjusted. For example, a more conservative forecast is provided to a user who is feeling stressed.

[0333] Results display

[0334] The server generates forecast results and inventory management proposals and sends them to the user's device. The device visually displays the results on a dashboard, allowing the user to intuitively understand the results. For example, graphs and tables are created using the Plotly library and provided to the user. The user can check the results and download the report in PDF format if necessary. The color and layout of the user interface are also dynamically adjusted based on the user's mood.

[0335] Example prompts to input to the generative AI model

[0336] "Please provide a sales forecast for October 2023. Please use the latest weather and economic indicator data. Please also take into account the user's emotional state and provide results."

[0337] In this way, the present invention enables sales forecasting that takes into account the emotional state of the user, thereby realizing a system that provides higher accuracy and a better user experience.

[0338] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0339] Step 1: Data collection

[0340] The server collects historical sales data, weather data, and economic indicator data.

[0341] Input: In-house databases, weather agency APIs, economic databases.

[0342] What it does: Uses Python to run SQL queries to extract sales data. Uses HTTP requests to retrieve weather data from meteorological agency APIs and economic indicators data from IMF and OECD APIs. Saves this data in CSV and JSON formats.

[0343] Output: Files with historical sales data, weather data, and economic indicator data.

[0344] Step 2: Data Preprocessing

[0345] The server pre-processes the collected data.

[0346] Inputs: Historical sales data, weather data, economic indicator data.

[0347] Specific operation: Uses the pandas library to impute missing values ​​in a data frame with the mean and remove outliers. Scales the data using scikit-learn's StandardScaler, which imputes missing values, removes noise, and scales the data.

[0348] Output: Preprocessed data.

[0349] Step 3: Model training

[0350] The server trains a machine learning model using the preprocessed data.

[0351] Input: Preprocessed data.

[0352] Specific operation: Create a model using scikit-learn's RandomForestRegressor, train the model using the training data, evaluate the model's accuracy using cross-validation, and adjust the hyperparameters as needed.

[0353] Output: A trained machine learning model.

[0354] Step 4: Emotion Recognition

[0355] The terminal recognizes the user's emotion data.

[0356] Input: User's facial expressions and vocal tone.

[0357] Specific operation: Analyzes video from a webcam using OpenCV and recognizes emotions from facial expressions. Extracts audio features using the Librosa library and classifies emotions using a deep learning model. Sends emotion data to the server in real time using WebSocket.

[0358] Output: Real-time sentiment data.

[0359] Step 5: Generate forecasts

[0360] The user submits a sales forecast request through a terminal.

[0361] Inputs: Sales forecast request from user, weather updates, economic indicators, and sentiment data.

[0362] How it works: The server receives the request and inputs the latest weather data, economic indicators, and sentiment data into the model. It then adjusts the prediction results to account for the user's emotional state and generates a sales forecast.

[0363] Output: Sales forecast results.

[0364] Step 6: View the results

[0365] The server generates the forecast results and inventory management proposals and sends them to the user's terminal.

[0366] Input: Sales forecast results, inventory management proposals.

[0367] Specific operation: The server generates forecast results and inventory management suggestions in JSON format and sends them to the terminal. The terminal uses the Plotly library to visually display them as graphs and tables on a dashboard. The UI color and layout are dynamically adjusted according to the user's mood. The user can also download the results in PDF format.

[0368] Output: Forecast results displayed in a dashboard, a report in PDF format.

[0369] (Application example 2)

[0370] 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."

[0371] Conventional sales forecasting systems make predictions based on past data and environmental data, but because they do not take user emotions into account, they often lack the accuracy of predictions and flexibility in customer service. Furthermore, in customer service situations in physical stores, there is no way to accurately grasp customer emotions, making it difficult to efficiently manage inventory and recommend appropriate products.

[0372] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting past sales data, weather data, and economic index data, means for preprocessing the collected data, imputing missing values, and normalizing the data, means for training a machine learning model based on the preprocessed data, means for recognizing user emotions in real time and using the emotion data, means for inputting the latest data and emotion data and generating a sales forecast in response to a sales forecast request from a user, and means for displaying the generated sales forecast results on the user's terminal and visually showing them on a dashboard. This enables flexible sales forecasting and inventory management based on user emotions, as well as customer service in physical stores.

[0373] "Historical sales data" refers to sales information recorded based on previous sales activity.

[0374] "Weather data" refers to numerical data or information that represents meteorological information, and is data that indicates the weather conditions at a specific date, time, and location.

[0375] "Economic indicator data" is data based on multiple economic indicators that show economic trends and conditions.

[0376] "Preprocessing" is the process of preparing collected data using methods such as filling in missing values, removing noise, and scaling.

[0377] A "machine learning model" is an algorithm that learns patterns and rules from large amounts of data and performs tasks such as prediction and classification.

[0378] "User emotion data" is information that indicates the user's emotional state in real time, obtained from the user's facial expressions, tone of voice, and the like.

[0379] An "emotion engine" is a technology or device that analyzes input data such as a user's facial expressions and voice to determine their emotional state.

[0380] A "sales forecast request" is a request made by a user to the system to forecast future sales.

[0381] "Latest data" refers to the most recent information or data available at the present time.

[0382] "Sales forecasting" refers to the use of machine learning models and other data to estimate future sales.

[0383] A "dashboard" is an interface designed for data visualization and analysis, and displays multiple pieces of information in an integrated manner.

[0384] "Inventory management proposal" is a method of proposing appropriate inventory levels and product placement based on past data and the latest information.

[0385] "PDF" is an abbreviation for Portable Document Format, a standard format for electronically storing and sharing documents.

[0386] This invention combines a system that collects past sales data, weather data, and economic indicator data and uses machine learning models to forecast sales, with an emotion engine that recognizes user emotions. This system is designed to improve sales efficiency and protect the environment, and also enables flexible responses based on user emotions.

[0387] The server comprises the following means:

[0388] 1. A means of collecting historical sales data, weather data, and economic indicator data

[0389] 2. Preprocessing the collected data, imputing missing values, and normalizing them

[0390] 3. A means to train machine learning models on preprocessed data

[0391] 4. Real-time recognition of user emotions and the use of emotional data

[0392] 5. A means of generating a sales forecast by inputting the latest data and sentiment data in response to a sales forecast request from a user.

[0393] 6. A means to display the generated sales forecast results on the user's device and visually display them on a dashboard

[0394] This will enable flexible sales forecasting and inventory management based on user sentiment, as well as customer service in physical stores.

[0395] Hardware and software used

[0396] Hardware

[0397] 1. Server: Performs key processing such as data collection, preprocessing, model training, and prediction generation.

[0398] 2. Smart glasses: Devices that detect the user's facial expressions and vocal tone to analyze emotions.

[0399] software

[0400] 1. OpenCV: An open-source library used for face detection and image processing.

[0401] 2. TensorFlow: A platform used to analyze facial expressions and recognize emotions using deep learning models.

[0402] 3. Scikit-Learn: A library used to perform sales forecasting using random forest regression models.

[0403] Example of a system

[0404] 1. Data Collection:

[0405] The server retrieves sales data for the past five years from an internal database, and also retrieves weather data (temperature, precipitation, etc.) for the past five years from meteorological agencies via API, as well as economic indicator data (GDP growth rate, unemployment rate).

[0406] 2. Data Preprocessing:

[0407] The server normalizes the collected data by imputing missing values, removing noise, and scaling. Missing sales data is imputed based on similar patterns in the past.

[0408] 3. Model training:

[0409] Using the pre-processed data, the server trains a random forest regression model to learn about past sales patterns and the effects of weather and economic conditions.

[0410] 4. Emotion recognition:

[0411] The smart glasses analyze the user's facial expressions and vocal tone in real time and use TensorFlow to obtain emotional data, which can determine whether the user is stressed or relaxed, for example.

[0412] 5. Sales forecast generation:

[0413] When a user requests "Sales forecast for October 2023," the server retrieves the latest weather forecast, economic indicator data, and sentiment data, and inputs these into the model to generate a sales forecast. If the server determines that the user is stressed, it will carefully provide feedback on the forecast and present the results.

[0414] 6. Results display:

[0415] The server sends the forecast results and inventory management suggestions to the user's device, which then visually displays them on a dashboard. The display content is also adjusted and the interface is customized based on the user's emotions.

[0416] Prompt Sentence Examples

[0417] "Local time in Tokyo on October 15, 2023, the weather is rainy, the temperature is 15 degrees, the GDP growth rate is 2.5%, the unemployment rate is 3%, and customer sentiment is stressed. Based on this data, generate a sales forecast."

[0418] This allows the present invention to leverage user sentiment and external data to optimize sales and customer experience in physical stores.

[0419] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0420] Step 1: Data collection

[0421] The server periodically collects historical sales data, weather data, and economic indicator data. The sales data is obtained from an internal company database, the weather data is obtained from a weather agency's API, and the economic indicator data is obtained from an economic database. The input is data from each data source, and the output is the collected data set.

[0422] Step 2: Data Preprocessing

[0423] The server preprocesses the collected data. Specifically, it imputes missing values ​​in the data, removes noise, and scales and normalizes it. For example, if there are gaps in sales data, it imputes them using similar data from the past. The input is the data collected in step 1, and the output is the preprocessed data.

[0424] Step 3: Model training

[0425] The server uses the preprocessed data to train a machine learning model. Training techniques such as random forest regression and LSTM (long short-term memory) networks are used. The model learns sales patterns based on past data. The input is the preprocessed data from step 2, and the output is the trained machine learning model.

[0426] Step 4: Emotion Recognition

[0427] The device analyzes the user's facial expressions and voice tone, and uses an emotion engine to recognize the user's emotions. The input is the user's voice and facial expression data acquired from the device, and the output is the analyzed emotion data.

[0428] Step 5: Generate a sales forecast

[0429] When a user sends a sales forecast request from their device, the server retrieves the latest weather and economic indicator data and inputs this data and sentiment data into the trained model. The server then adjusts the forecast and generates a sales forecast. The input is the user's request, the latest data, and sentiment data, and the output is the sales forecast result.

[0430] Step 6: View the results

[0431] The server generates forecast results and inventory management proposals and sends them to the user's terminal, where they are visually displayed as a dashboard. The input is the sales forecast results, and the output is the visually displayed data.

[0432] This process flow allows the present invention to respond to user emotions while providing more accurate sales forecasts and inventory management.

[0433] 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.

[0434] 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.

[0435] 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.

[0436] [Second embodiment]

[0437] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0438] 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.

[0439] 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).

[0440] 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.

[0441] 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.

[0442] 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).

[0443] 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.

[0444] 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.

[0445] 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.

[0446] 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.

[0447] In the smart glasses 214, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0448] 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."

[0449] The present invention is a system that collects past sales data, weather data, and economic indicator data, and uses machine learning models to forecast sales based on this data. This system is designed to improve sales efficiency and protect the environment, and an embodiment of the system is described below.

[0450] System Overview

[0451] This system consists of a server, a user's device, and a network infrastructure that connects them. The server has multiple functions, such as data collection, preprocessing, model training, prediction generation, and result display. Meanwhile, users input requests through their device and check the prediction results.

[0452] Program processing flow

[0453] 1. Data Collection

[0454] The server periodically collects sales data, weather data, and economic indicator data. The sales data is obtained from an internal database, the weather data is obtained from a weather agency's API, and the economic indicator data is obtained from an economic database.

[0455] 2. Data Preprocessing

[0456] The server preprocesses the collected data by imputing missing values, removing noise, scaling, and normalizing it. For example, for periods where sales data is missing, it uses similar data from the past to impute the missing data.

[0457] 3. Model Training

[0458] The server uses the preprocessed data to train a machine learning model, which uses a selected algorithm (e.g., random forest regression or LSTM) for sales forecasting. The model learns sales patterns based on past data.

[0459] 4. Prediction Generation

[0460] When a user sends a sales forecast request from their device, the server retrieves the latest weather data and economic trend data from the API and uses this data to generate a sales forecast in real time.

[0461] 5. Display results

[0462] The server generates the prediction results and sends them to the user's device, where they can view the results visually on a dashboard and optionally download the report in PDF format.

[0463] Specific examples

[0464] Data collection

[0465] The server retrieves sales data for the past five years from an internal database, weather data (temperature, precipitation, etc.) for the past five years from the Japan Meteorological Agency via an API, and economic indicator data (e.g., GDP growth rate, unemployment rate) from the government and reliable economic forums.

[0466] Data Preprocessing

[0467] The server fills in missing data from the collected sales data using past approximate data. It also scales the sales data and weather data to a range of 0 to 1 and converts them into a format suitable for the model.

[0468] Model learning

[0469] The server trains a random forest regression model on the preprocessed data, which is used to learn about past sales patterns and the effects of weather and economic conditions and to predict future sales.

[0470] Forecast Generation

[0471] When a user requests "Sales forecast for October 2023," the server retrieves the latest weather forecast and economic indicators and inputs them into the trained model, which then outputs detailed sales forecasts and supply and demand plans for each region.

[0472] Results display

[0473] The server organizes the generated sales forecast results and displays them on the user's dashboard, where users can view the data in graphs and tables and download the reports in PDF format if required.

[0474] This frees sales representatives from manual work, enabling more efficient sales activities and achieving sustainable inventory management across the entire company.

[0475] The processing flow will be explained below.

[0476] Step 1:

[0477] The server retrieves past sales data from an internal database, which stores information such as sales date, product ID, sales quantity, price, and area.

[0478] Step 2:

[0479] The server obtains weather data from the weather agency's API, including past temperature, precipitation, wind speed, and air pressure.

[0480] Step 3:

[0481] The server retrieves economic indicator data from economic databases or trusted government APIs, such as GDP growth, unemployment rate, and consumer confidence index.

[0482] Step 4:

[0483] The server aggregates the collected sales, weather, and economic data, making the data set available in a consistent manner.

[0484] Step 5:

[0485] The server preprocesses the integrated data, imputing missing values, removing noise, and scaling the data. For example, if there is missing sales data, it is imputed using similar data from the past.

[0486] Step 6:

[0487] The server normalizes the data by scaling each variable to a range between 0 and 1. This step is important for improving the accuracy of training machine learning models.

[0488] Step 7:

[0489] The server uses the preprocessed data to train machine learning models, for example, using random forest regression or LSTM (long short-term memory) networks.

[0490] Step 8:

[0491] The server uses validation data to optimize model parameters during training and prevent overfitting or underfitting, a process that evaluates how well the model performs on real-world data.

[0492] Step 9:

[0493] The server compares the predictions with actual sales data to evaluate the accuracy of the model and tunes the model as needed, a step that is important for continuously improving the model's accuracy.

[0494] Step 10:

[0495] A user enters a request into a terminal, such as "Sales forecast for October 2023," which starts the forecast generation process.

[0496] Step 11:

[0497] The server retrieves the latest weather and economic indicator data in real time and inputs this into the trained model.

[0498] Step 12:

[0499] The server generates a sales forecast, which includes projected sales figures, peak demand times, and geographical details.

[0500] Step 13:

[0501] The server also generates inventory management recommendations based on the generated forecasts, helping to prevent overstocks and out-of-stock situations.

[0502] Step 14:

[0503] The server sends the prediction results and inventory management suggestions to the user's terminal.

[0504] Step 15:

[0505] The device visually displays the received prediction results on a dashboard, allowing users to view the results in graph and table format.

[0506] Step 16:

[0507] Users can check the forecast results on the dashboard and download the report in PDF format as needed, which allows them to develop specific sales strategies and inventory management plans.

[0508] As described above, the system collects data, preprocesses it, trains models, generates predictions, and displays the results step by step, supporting efficient sales activities and sustainable inventory management.

[0509] Example 1

[0510] 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."

[0511] Conventional sales forecasting systems often make predictions based solely on past sales data, without taking into account multiple factors such as weather data and economic indicators, resulting in low accuracy. Another issue is that few systems provide sales forecast results in real time, and the effects of missing data and noise cannot be fully mitigated. Furthermore, there is a lack of functionality that allows users to easily check sales forecast results and provides them in a format that can be used as reference material. These issues make it difficult to improve sales efficiency and protect the environment.

[0512] 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.

[0513] In this invention, the server includes: means for periodically collecting past sales data, weather data, and economic indicator data from an internal company database or an external API; means for preprocessing the collected data, filling in missing values ​​with past approximate data, removing noise, scaling, and normalizing the data; means for training a machine learning model based on the preprocessed data to learn past sales patterns and environmental factors; means for inputting the latest weather data and economic indicator data in response to a sales forecast request from a user and generating a sales forecast in real time; and means for transmitting the generated sales forecast results to the user's device, visually displaying them on a dashboard, and making them available for download as a PDF report if necessary. This improves the accuracy of sales forecasts and enables real-time information provision and visual display of results. As a result, sales efficiency and sustainable inventory management can be achieved.

[0514] An "in-house database" is a system installed within a company that stores and manages information, and holds business data such as sales data and customer information.

[0515] An "external organization's API" is an API (Application Programming Interface) provided by an external organization or service, and is a means of programmatically obtaining information such as weather data and economic indicator data.

[0516] "Sales data" refers to numerical information about sales recorded by a company during a specific period, such as the sales quantity and sales amount of a product.

[0517] "Weather data" is numerical information about the weather in a specific region or period, such as temperature, precipitation, humidity, and wind speed.

[0518] "Economic indicator data" refers to various statistical data that show the state of the economy, such as GDP growth rate, unemployment rate, and consumer price index.

[0519] "Preprocessing" refers to the preparation of data prior to data analysis or machine learning, and specifically includes processes such as filling in missing values, removing noise, scaling, and normalization.

[0520] "Missing values" refers to a state in which some data is missing, including blanks and NULL values ​​in a dataset.

[0521] "Noise" refers to unnecessary disturbances and errors contained in data that can reduce the accuracy of analyses and models.

[0522] "Scaling" is a conversion technique used to fit data values ​​into a range, usually normalizing the data range to 0 to 1 using minimum and maximum values.

[0523] "Normalization" refers to a transformation process to make the distribution of data uniform, particularly by confining the data to a specific range through scaling.

[0524] A "machine learning model" is a set of algorithms that learn patterns and rules from data and make predictions or classifications. Examples include random forests and LSTM.

[0525] "Training" is the process by which a machine learning model uses data to learn optimal parameters.

[0526] A "sales forecast request" is a request sent by a user to a server to forecast future sales based on a specific period and conditions.

[0527] "Real-time" refers to a state in which data is processed and displayed immediately at the moment it is generated or acquired.

[0528] A "dashboard" is a visual interface that allows users to see important information at a glance, displaying data in the form of graphs and tables.

[0529] "PDF format" stands for Portable Document Format, a file format for storing and sharing information in a fixed-layout format.

[0530] A "report" refers to the analysis results or report on specific data, and is often generated in PDF format or similar.

[0531] This invention is a system that collects past sales data, weather data, and economic indicator data from in-house databases and APIs of external institutions, and uses this data to perform sales forecasts using machine learning models. This system consists of a server, user terminals, and a network infrastructure that connects these.

[0532] Hardware and Software Configuration

[0533] server

[0534] The server is a computer system equipped with a high-performance processor and large memory capacity, and is installed with software for database management systems, API handlers, and machine learning models, including pandas, scikit-learn, and TensorFlow for data analysis.

[0535] Terminal

[0536] Users access the system through a web browser on devices such as PCs and tablets. The device must have an internet connection and a browser (e.g., Google Chrome or Mozilla Firefox) to view dashboards and reports.

[0537] Data collection

[0538] The server periodically connects to the company's internal database and retrieves sales data for the past five years, which is retrieved using SQL queries.

[0539] The server retrieves weather data for the past five years through the API of an external weather agency. This API request uses the HTTP GET method to retrieve data such as temperature and precipitation for a specified period.

[0540] The server uses APIs provided by governments and economic forums to collect economic indicator data (e.g., GDP growth rate, unemployment rate) for the past five years.

[0541] Data Preprocessing

[0542] The server uses the pandas library to fill in missing values ​​in the acquired sales data with similar historical data, and applies appropriate filtering techniques to remove noise.

[0543] Next, we use scikit-learn's MinMaxScaler to scale and normalize the sales and weather data to the range 0 to 1, so that the data can be fed into the machine learning model in a unified format.

[0544] Model learning

[0545] The server trains a machine learning model based on the preprocessed data. Algorithms such as random forest regression models and LSTM are used, and the model is run using scikit-learn and TensorFlow. During the training process, the dataset is split into training data and validation data, and cross-validation is applied to ensure the accuracy of the model.

[0546] Forecast Generation

[0547] When a user submits a sales forecast request from their device, they enter a prompt, such as "Sales forecast for October 2023." In response, the server again retrieves the latest weather and economic data and inputs it into the trained model.

[0548] The model generates results by predicting future sales in real time based on new data.

[0549] Results display

[0550] The server organizes the generated prediction results and sends them to the user's device, where they are visually displayed in graphs and tables on a dashboard for the user to review.

[0551] If desired, users can download the report in PDF format, a feature that uses libraries such as matplotlib and ReportLab.

[0552] Specific examples

[0553] Data collection

[0554] The server retrieves sales data for the past five years from a company database, collects weather data from the Japan Meteorological Agency's API, and obtains economic indicator data from the government's economic forum.

[0555] Data Preprocessing

[0556] The server fills in missing sales data with similar data from the past and scales the weather data to normalize it to the range 0 to 1.

[0557] Model learning

[0558] The server trains a random forest regression model to learn patterns from the data.

[0559] Forecast Generation

[0560] When a user requests "Sales forecast for October 2023," the server inputs the latest data into the model and generates a sales forecast.

[0561] Results display

[0562] The server displays the prediction results on a dashboard and allows users to download the report in PDF format.

[0563] Prompt Sentence Examples

[0564] "Show me sales forecast for October 2023."

[0565] "Get weather data for the past five years."

[0566] "Please update your economic indicator data."

[0567] This system frees sales representatives from manual work, enabling more efficient sales activities and achieving sustainable inventory management across the entire company.

[0568] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0569] Step 1:

[0570] The server connects to the company's internal database and retrieves sales data for the past five years. This is done using an SQL query. Specifically, the server executes a query such as "SELECT FROM sales WHERE date BETWEEN '2018-01-01' AND '2022-12-31'" to extract the required data. The input is the connection information and the sales data contained in the database, and the output is the sales data for the specified period.

[0571] Step 2:

[0572] The server sends an HTTP GET request to the weather agency's API to retrieve weather data for the past five years. Specifically, it sends a request such as "GET http: / / api.weather.com / data?startDate=2018-01-01&endDate=2022-12-31" to retrieve temperature and precipitation data. The input is the API endpoint information, and the output is the weather data returned by the API.

[0573] Step 3:

[0574] The server accesses the government's economic data API to obtain economic indicator data (e.g., GDP growth rate and unemployment rate) for the past five years. Specifically, it executes "GET http: / / api.economicdata.com / indicators?startDate=2018-01-01&endDate=2022-12-31" to obtain the indicator data. The input is the API endpoint information, and the output is the obtained economic indicator data.

[0575] Step 4:

[0576] The server preprocesses the collected sales data, weather data, and economic indicator data. Specifically, it creates a data frame using the pandas library and imputes missing values ​​using methods such as linear imputation. The input is the collected raw data, and the output is a data frame with missing values ​​imputed.

[0577] Step 5:

[0578] The server scales and normalizes the preprocessed data, specifically using scikit-learn's MinMaxScaler to convert values ​​to the range 0 to 1. The input is the interpolated data frame, and the output is the scaled data.

[0579] Step 6:

[0580] The server uses the scaled data to train a machine learning model, for example, a random forest regression model using scikit-learn's RandomForestRegressor. The input is the preprocessed and scaled data, and the output is a trained machine learning model.

[0581] Step 7:

[0582] The user sends a sales forecast request from the terminal. For example, they input a prompt such as "Sales forecast for October 2023." The input is the user's request, and the output is the request data to the server.

[0583] Step 8:

[0584] The server retrieves the latest weather data and economic indicator data from the API again and inputs it into the trained model, generating a sales forecast in real time. The input is the latest weather data and economic indicator data, and the output is sales forecast data.

[0585] Step 9:

[0586] The server organizes the generated sales forecast results and sends them to the user's device. The data is displayed visually on a dashboard. The input is the sales forecast data, and the output is the graphs and tables displayed on the dashboard.

[0587] Step 10:

[0588] Users can check the forecast results through a dashboard and download them in PDF format if necessary. The server generates reports using libraries such as matplotlib and ReportLab. The input is the user's download request and sales forecast data, and the output is a PDF report.

[0589] (Application example 1)

[0590] 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."

[0591] In traditional manufacturing, demand forecasting and inventory management are often performed manually, resulting in problems such as human error and reduced efficiency. It is also difficult to flexibly adjust manufacturing processes based on real-time data from the field, making it difficult to maintain a balance between supply and demand. Furthermore, it is difficult to make accurate predictions that take environmental fluctuations (weather and economic conditions) into account, leading to inventory surpluses and shortages. This calls for a method to improve production efficiency, minimize resource waste, and establish a sustainable production system.

[0592] 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.

[0593] In this invention, the server includes: means for collecting past sales data, weather data, and economic index data; means for preprocessing the collected data, imputing missing values, and normalizing the data; means for training a machine learning model based on the preprocessed data; means for inputting the latest data and generating a sales forecast in response to a sales forecast request from a user; means for displaying the generated sales forecast results on the user's terminal and visually showing them on a dashboard; and means for instructing a robot to adjust the manufacturing process based on the generated forecast results. This allows the factory's manufacturing process to be automatically adjusted in real time based on supply and demand forecasts, not only improving production efficiency but also reducing human error and enabling sustainable inventory management.

[0594] "Past sales data" refers to sales performance information recorded by companies and stores to date, specifically information such as product sales quantity, sales amount, and sales period.

[0595] "Weather data" refers to meteorological information such as temperature, precipitation, and wind speed obtained from meteorological agencies and used in forecasting models.

[0596] "Economic indicator data" refers to data that shows economic trends, and includes GDP growth rate, unemployment rate, consumer price index, etc.

[0597] "Means of collection" refers to the methods and technologies used to obtain the necessary information from various data sources.

[0598] "Preprocessing means" refers to methods and techniques for converting acquired data into an analyzable format, and specifically includes imputing missing values ​​and normalizing data.

[0599] "Methods for imputing missing values" are methods for completing incomplete data by using similar data from the past or average values, etc.

[0600] "Normalization" is a method of scaling data values ​​into a certain range so that machine learning algorithms can learn efficiently.

[0601] A "means for training a machine learning model" is a method for training a machine learning algorithm using preprocessed data to build a predictive model.

[0602] The "means for generating sales forecasts" is a method for inputting the latest data into a trained machine learning model to predict future sales.

[0603] A "user's terminal" is a device such as a PC or smartphone that the user uses to check the prediction results.

[0604] A "dashboard" refers to an interface that allows users to visually check prediction results, displaying information in the form of graphs and tables.

[0605] "Means for issuing instructions to robots to adjust the manufacturing process based on the generated prediction results" refers to methods and technologies for issuing specific operating instructions to manufacturing robots based on predicted supply and demand data, automatically adjusting the manufacturing line.

[0606] The present invention is a system that collects past sales data, weather data, and economic indicator data, and uses a machine learning model based on this data to forecast sales, and this system has been applied to a factory robot. This system is configured as follows.

[0607] System Overview

[0608] This system consists of a server, user terminals, factory robots, and a network infrastructure that connects them. The server has multiple functions, including data collection, preprocessing, model learning, prediction generation, and result display. Meanwhile, users input requests through terminals or consoles and check prediction results. Furthermore, the factory robots automatically adjust the manufacturing process based on this prediction data.

[0609] Hardware and software used

[0610] Hardware:

[0611] Factory robots (e.g. Fanuc, ABB, KUKA, etc.)

[0612] Servers (cloud or on-site data center)

[0613] Sensors (monitor temperature, humidity, vibration, etc.)

[0614] software:

[0615] Machine learning frameworks (e.g. TensorFlow, PyTorch)

[0616] Database (e.g. MySQL, PostgreSQL)

[0617] API services (e.g. OpenWeatherMap, economic data API)

[0618] Robot control software (e.g. ROS - Robot Operating System)

[0619] Data processing and calculation

[0620] 1. The server retrieves past sales data from the company's internal database, collects weather data from the weather agency's API, and collects economic indicator data from the economic database.

[0621] 2. The server preprocesses the collected data. Specifically, it imputes missing values, removes noise, and scales and normalizes the data. For example, if sales data is missing for a period, it is imputed using similar data from the past.

[0622] 3. The server uses the preprocessed data to train a machine learning model. This model uses a selected algorithm (e.g., random forest regression or LSTM) for sales forecasting. The model learns sales patterns based on past data.

[0623] 4. When a user sends a sales forecast request from their device, the server retrieves the latest weather data and economic trend data from the API and uses this data to generate a sales forecast in real time.

[0624] 5. The server generates the prediction results and sends them to the user's device. The user can view the results visually on a dashboard and optionally download them as a PDF report.

[0625] 6. Based on the generated prediction results, the server sends instructions to the factory robot to adjust the manufacturing process, allowing the factory robot to automatically optimize the manufacturing process based on the supply and demand balance.

[0626] Specific examples

[0627] Data collection:

[0628] The server retrieves the past two years of production and inventory data from the factory database, and also uses APIs to retrieve the past two years of weather data (e.g., temperature, humidity) and economic data (e.g., GDP growth rate, production index).

[0629] Data preprocessing:

[0630] The server fills in missing data using past approximate data, scales sales data and weather data to a range of 0 to 1, and converts them into a format suitable for the model.

[0631] Model training:

[0632] The server trains a random forest regression model on the preprocessed data, which is used to learn about past sales patterns and the effects of weather and economic conditions and to predict future sales.

[0633] Forecast generation:

[0634] When a user requests "Sales forecast for October 2023," the server retrieves the latest weather forecast and economic indicators and inputs them into the trained model, which then outputs detailed sales forecasts and supply and demand plans for each region.

[0635] Results display and automatic adjustment:

[0636] The server then organizes the generated sales forecasts and displays them on the user's dashboard. Based on the forecasts, the server sends instructions to factory robots to adjust the manufacturing process, for example, adjusting the production speed of a specific product line to prevent overstocks or shortages.

[0637] Example prompts to input to the generative AI model

[0638] Input historical production data, weather data, and economic indicator data to generate a production demand forecast for the next month.

[0639] Using this system, supply and demand forecasts and production process optimization at manufacturing sites can be performed in real time, improving production efficiency and optimizing inventory management.

[0640] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0641] Step 1: Data collection

[0642] The server retrieves past sales data from the company's internal database, and uses APIs to collect weather data from meteorological agencies and economic indicator data from economic databases. As a result, past sales data, weather data, and economic indicator data are aggregated on the server as input data. Specifically, this includes sales quantity, sales amount, temperature, precipitation, GDP growth rate, unemployment rate, etc.

[0643] Step 2: Data Preprocessing

[0644] The server then fills in missing values ​​in the collected data, removes noise, and scales and normalizes the data. Specifically, it fills in missing sales data using similar historical data, and does the same for weather data and economic indicator data. The result is a preprocessed dataset that can be used by machine learning models.

[0645] Step 3: Train the machine learning model

[0646] The server uses the preprocessed data to train a machine learning model for sales forecasting. Specifically, it uses algorithms such as random forest regression and LSTM to learn sales patterns and correlations with various metrics, resulting in a trained machine learning model.

[0647] Step 4: Receive a prediction request

[0648] The user sends a sales forecast request from the terminal. For example, they enter a prompt sentence such as "Please generate sales forecast for October 2023." This becomes the input data to the server.

[0649] Step 5: Generate forecasts

[0650] The server receives the request from the user and retrieves the latest weather and economic indicator data from the API again. The retrieved data is input into the trained model to generate a sales forecast. This operation outputs real-time sales forecast data.

[0651] Step 6: View the results

[0652] The server sends the generated sales forecast results to the user's device, where the user can visually check the forecast results on the dashboard of their device. For example, the sales forecast data can be displayed in graph or table format.

[0653] Step 7: Adjust the manufacturing process based on the results

[0654] Based on the generated predictions, the server sends instructions to the factory robots to adjust the manufacturing process. The factory robots then use the instructions received from the server to optimize the manufacturing process in real time, for example, by adjusting the production speed of a particular product line or changing the amount of materials required.

[0655] Through these processing steps, the system automates everything from sales forecasts to real-time adjustments to manufacturing processes, enabling efficient production management.

[0656] 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.

[0657] This invention combines a system that collects past sales data, weather data, and economic indicator data and uses machine learning models to forecast sales, with an emotion engine that recognizes user emotions. This system is designed to improve sales efficiency and protect the environment, and also enables flexible responses based on user emotions.

[0658] System Overview

[0659] This system consists of a server, a user's device, an emotion engine, and a network infrastructure that connects these components. The server has multiple functions, including data collection, preprocessing, model training, prediction generation, result display, and emotion recognition. Meanwhile, users input requests through their device and check prediction results.

[0660] Program processing flow

[0661] 1. Data Collection

[0662] The server periodically collects historical sales data, weather data, and economic indicator data. The sales data is obtained from an internal database, the weather data is obtained from a weather agency's API, and the economic indicator data is obtained from an economic database.

[0663] 2. Data Preprocessing

[0664] The server preprocesses the collected data by filling in missing values, removing noise, scaling, and normalizing it. For example, if there are gaps in sales data, it fills them in using similar data from the past.

[0665] 3. Model Training

[0666] The server uses the preprocessed data to train a machine learning model using techniques such as random forest regression and LSTM (long short-term memory) networks. The model learns sales patterns based on past data.

[0667] 4. Emotion recognition

[0668] The device analyzes the user's facial expressions and tone of voice, and uses an emotion engine to recognize the user's emotions. The emotion data is sent to the server in real time.

[0669] 5. Prediction Generation

[0670] When a user sends a sales forecast request from their device, the server retrieves the latest weather and economic indicator data, inputs this data into the trained model, and adjusts the forecast taking into account emotion data from the emotion engine.

[0671] 6. Results display

[0672] The server generates prediction results and inventory management suggestions and sends them to the user's device. The user can view the results visually on a dashboard, and emotion data is used to dynamically adjust the display content and interface.

[0673] Specific examples

[0674] Data collection

[0675] The server retrieves sales data for the past five years from an internal database, and also retrieves weather data (temperature, precipitation, etc.) for the past five years from meteorological agencies via API, as well as economic indicator data (GDP growth rate, unemployment rate).

[0676] Data Preprocessing

[0677] The server normalizes the collected data by imputing missing values, removing noise, and scaling. Missing sales data is interpolated based on similar patterns in the past.

[0678] Model learning

[0679] The server trains a random forest regression model on the pre-processed data to learn about past sales patterns and the effects of weather and economic conditions.

[0680] emotion recognition

[0681] The device analyzes the user's facial expressions and voice tone in real time and obtains emotional data through an emotion engine, which determines whether the user is, for example, stressed or relaxed.

[0682] Forecast Generation

[0683] When a user requests "Sales forecast for October 2023," the server retrieves the latest weather forecast, economic indicator data, and sentiment data, and inputs these into the model to generate a sales forecast. If the server determines that the user is stressed, it will carefully provide feedback on the forecast and present the results.

[0684] Results display

[0685] The server sends the forecast results and inventory management suggestions to the user's device, which displays them in graph and table format on a dashboard. The user can check the results and download the report in PDF format if necessary. The display content is also adjusted and the interface is customized based on the user's sentiment.

[0686] In this way, the present invention can provide more accurate sales forecasts and inventory management while responding to user emotions.

[0687] The processing flow will be explained below.

[0688] Step 1:

[0689] The server retrieves past sales data from an internal database, which stores information such as sales date, product ID, sales quantity, price, and area.

[0690] Step 2:

[0691] The server obtains weather data from the weather agency's API, specifically, past temperature, precipitation, wind speed, air pressure, etc.

[0692] Step 3:

[0693] The server retrieves economic indicator data from economic databases or APIs of trusted government agencies, such as GDP growth, unemployment rate, and consumer confidence index.

[0694] Step 4:

[0695] The server aggregates the collected sales data, weather data, and economic indicator data, allowing for a consistent analysis of the data set.

[0696] Step 5:

[0697] The server preprocesses the integrated data by imputing missing values, removing noise, and scaling it. For example, if there is missing sales data, it fills in the missing data using similar data from the past.

[0698] Step 6:

[0699] The server uses the preprocessed data as training data for machine learning models, such as random forest regression and LSTM (long short-term memory) networks, to train the models.

[0700] Step 7:

[0701] The server uses validation data to optimize the parameters of the trained model and prevent overfitting and underfitting.

[0702] Step 8:

[0703] The device analyzes the user's facial expressions and voice tone in real time and recognizes their emotions using an emotion engine. The recognized emotion data is then sent to the server.

[0704] Step 9:

[0705] When a user requests "Sales forecast for October 2023" from their device, the server retrieves the latest weather data, economic indicator data, and sentiment data.

[0706] Step 10:

[0707] The server then inputs this data into a trained model to generate sales forecasts, with the sentiment data being used to fine-tune the results.

[0708] Step 11:

[0709] The server organizes the generated sales forecast results and simultaneously creates inventory management proposals, thereby preventing excess inventory and out-of-stock situations.

[0710] Step 12:

[0711] The server transmits the sales forecast results and inventory management proposals to the user's terminal.

[0712] Step 13:

[0713] The terminal displays the received sales forecast results and inventory management suggestions on a dashboard, including graphs and tables.

[0714] Step 14:

[0715] Users can check the forecast results on the dashboard and download the report in PDF format if necessary.

[0716] Step 15:

[0717] The device dynamically adjusts the display content and interface based on the user's emotions. For example, if the user is feeling stressed, the device will change the display to a simpler display and a more visually soothing color scheme.

[0718] As described above, the system performs data collection, preprocessing, model training, emotion recognition, prediction generation, and result display at each step, supporting efficient sales activities and sustainable inventory management while also improving the user experience.

[0719] Example 2

[0720] 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."

[0721] Conventional sales forecasting systems typically collect past sales data, weather data, economic indicator data, etc., and use machine learning to forecast sales. However, they were unable to take into account individual factors such as the user's emotional state. This meant that they were unable to respond flexibly to user emotions, making it difficult to improve forecast accuracy and user experience. Furthermore, there were limited ways to visually confirm forecast results and present them in a format that was easy for users to understand.

[0722] 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.

[0723] In this invention, the server includes: means for collecting past sales data, weather data, and economic indicator data; means for preprocessing the collected data, imputing missing values, and normalizing the data; means for training a machine learning model based on the preprocessed data; means for recognizing user emotion data and transmitting it to the server; means for inputting the latest data and emotion data and generating a sales forecast in response to a sales forecast request from the user; and means for displaying the generated sales forecast results on the user's terminal and visually presenting them on a dashboard. This enables sales forecasts that take the user's emotional state into account, improving forecast accuracy and the user experience. Furthermore, visually displaying the forecast results makes it possible to provide a system that is easy for users to understand and use.

[0724] "Past sales data" refers to records of past sales and purchases of goods conducted by a company, and includes data such as sales quantities, amounts, and dates.

[0725] "Weather data" refers to data including meteorological information such as temperature, precipitation, wind speed, and humidity obtained from meteorological agencies.

[0726] "Economic indicator data" is statistical data used to understand the state of the economy, such as gross domestic product (GDP), unemployment rate, and inflation rate.

[0727] "Preprocessing" refers to the processing performed on collected raw data, and involves cleansing and normalizing the data, such as filling in missing values, removing noise, and scaling.

[0728] A "machine learning model" is an algorithm or computational model that learns patterns and regularities from given data and makes predictions and judgments about future data.

[0729] "Emotion data" is data that indicates the emotional state of the user analyzed from facial expressions, tone of voice, and the like.

[0730] A "server" is a central computer that performs data collection, preprocessing, model training, prediction generation, and display of results.

[0731] A "terminal" is a computer or mobile device used by a user to enter data and confirm results.

[0732] A "dashboard" is an interface that allows users to visually check information such as sales forecast results and inventory management proposals.

[0733] This invention is a system that collects and analyzes past sales data, weather data, economic indicator data, and even user emotion data, and uses machine learning models to forecast sales. This system is composed of a server, a terminal, an emotion recognition engine, and a network infrastructure that links these components. Below, we will explain in detail how each component works.

[0734] Data collection

[0735] The server periodically collects past sales data, weather data, and economic indicator data. Sales data is obtained from an internal database, weather data is obtained using the weather agency's API, and economic indicator data is obtained from an economic database. Specifically, sales data is extracted from the database by executing SQL queries using Python. Weather data is obtained in JSON format from the weather agency's API using an HTTP request, for example, and economic indicator data is collected from the IMF and OECD APIs.

[0736] Data Preprocessing

[0737] The server preprocesses the collected data. This preprocessing includes imputing missing values, removing noise, scaling, and normalizing the data. For example, it uses Python's pandas library to impute missing values ​​in the data frame with the mean value and scales the data using scikit-learn's StandardScaler. This cleans the data so that the machine learning model can operate optimally.

[0738] Model learning

[0739] The server uses the preprocessed data to train a machine learning model. The main techniques used are random forest regression and LSTM (long short-term memory). For example, a model is created using scikit-learn's RandomForestRegressor and trained using the training data. The server also evaluates the accuracy of the model and adjusts hyperparameters using cross-validation if necessary.

[0740] emotion recognition

[0741] The device captures the user's facial expressions and voice tone and analyzes them with an emotion recognition engine. For example, it uses OpenCV to analyze video from a webcam and recognize emotions from facial expressions. It also uses the Librosa library to extract voice features and classify emotions using a deep learning model. This emotion data is sent to a server in real time and incorporated into a prediction algorithm.

[0742] Forecast Generation

[0743] When a user requests a sales forecast through their device, the server retrieves the latest weather data, economic indicator data, and emotion data and inputs this into the trained model. For example, if a user requests "Sales forecast for October 2023," the server collects the data and generates a sales forecast. The user's emotion data is also taken into account and the forecast results are adjusted. For example, a more conservative forecast is provided to a user who is feeling stressed.

[0744] Results display

[0745] The server generates forecast results and inventory management proposals and sends them to the user's device. The device visually displays the results on a dashboard, allowing the user to intuitively understand the results. For example, graphs and tables are created using the Plotly library and provided to the user. The user can check the results and download the report in PDF format if necessary. The color and layout of the user interface are also dynamically adjusted based on the user's mood.

[0746] Example prompts to input to the generative AI model

[0747] "Please provide a sales forecast for October 2023. Please use the latest weather and economic indicator data. Please also take into account the user's emotional state and provide results."

[0748] In this way, the present invention enables sales forecasting that takes into account the emotional state of the user, thereby realizing a system that provides higher accuracy and a better user experience.

[0749] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0750] Step 1: Data collection

[0751] The server collects historical sales data, weather data, and economic indicator data.

[0752] Input: In-house databases, weather agency APIs, economic databases.

[0753] What it does: Uses Python to run SQL queries to extract sales data. Uses HTTP requests to retrieve weather data from meteorological agency APIs and economic indicators data from IMF and OECD APIs. Saves this data in CSV and JSON formats.

[0754] Output: Files with historical sales data, weather data, and economic indicator data.

[0755] Step 2: Data Preprocessing

[0756] The server pre-processes the collected data.

[0757] Inputs: Historical sales data, weather data, economic indicator data.

[0758] Specific operation: Uses the pandas library to impute missing values ​​in a data frame with the mean and remove outliers. Scales the data using scikit-learn's StandardScaler, which imputes missing values, removes noise, and scales the data.

[0759] Output: Preprocessed data.

[0760] Step 3: Model training

[0761] The server trains a machine learning model using the preprocessed data.

[0762] Input: Preprocessed data.

[0763] Specific operation: Create a model using scikit-learn's RandomForestRegressor, train the model using the training data, evaluate the model's accuracy using cross-validation, and adjust the hyperparameters as needed.

[0764] Output: A trained machine learning model.

[0765] Step 4: Emotion Recognition

[0766] The terminal recognizes the user's emotion data.

[0767] Input: User's facial expressions and vocal tone.

[0768] Specific operation: Analyzes video from a webcam using OpenCV and recognizes emotions from facial expressions. Extracts audio features using the Librosa library and classifies emotions using a deep learning model. Sends emotion data to the server in real time using WebSocket.

[0769] Output: Real-time sentiment data.

[0770] Step 5: Generate forecasts

[0771] The user submits a sales forecast request through a terminal.

[0772] Inputs: Sales forecast request from user, weather updates, economic indicators, and sentiment data.

[0773] How it works: The server receives the request and inputs the latest weather data, economic indicators, and sentiment data into the model. It then adjusts the prediction results to account for the user's emotional state and generates a sales forecast.

[0774] Output: Sales forecast results.

[0775] Step 6: View the results

[0776] The server generates the forecast results and inventory management proposals and sends them to the user's terminal.

[0777] Input: Sales forecast results, inventory management proposals.

[0778] Specific operation: The server generates forecast results and inventory management suggestions in JSON format and sends them to the terminal. The terminal uses the Plotly library to visually display them as graphs and tables on a dashboard. The UI color and layout are dynamically adjusted according to the user's mood. The user can also download the results in PDF format.

[0779] Output: Forecast results displayed in a dashboard, a report in PDF format.

[0780] (Application example 2)

[0781] 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."

[0782] Conventional sales forecasting systems make predictions based on past data and environmental data, but because they do not take user emotions into account, they often lack the accuracy of predictions and flexibility in customer service. Furthermore, in customer service situations in physical stores, there is no way to accurately grasp customer emotions, making it difficult to efficiently manage inventory and recommend appropriate products.

[0783] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting past sales data, weather data, and economic index data, means for preprocessing the collected data, imputing missing values, and normalizing the data, means for training a machine learning model based on the preprocessed data, means for recognizing user emotions in real time and using the emotion data, means for inputting the latest data and emotion data and generating a sales forecast in response to a sales forecast request from a user, and means for displaying the generated sales forecast results on the user's terminal and visually showing them on a dashboard. This enables flexible sales forecasting and inventory management based on user emotions, as well as customer service in physical stores.

[0784] "Historical sales data" refers to sales information recorded based on previous sales activity.

[0785] "Weather data" refers to numerical data or information that represents meteorological information, and is data that indicates the weather conditions at a specific date, time, and location.

[0786] "Economic indicator data" is data based on multiple economic indicators that show economic trends and conditions.

[0787] "Preprocessing" is the process of preparing collected data using methods such as filling in missing values, removing noise, and scaling.

[0788] A "machine learning model" is an algorithm that learns patterns and rules from large amounts of data and performs tasks such as prediction and classification.

[0789] "User emotion data" is information that indicates the user's emotional state in real time, obtained from the user's facial expressions, tone of voice, and the like.

[0790] An "emotion engine" is a technology or device that analyzes input data such as a user's facial expressions and voice to determine their emotional state.

[0791] A "sales forecast request" is a request made by a user to the system to forecast future sales.

[0792] "Latest data" refers to the most recent information or data available at the present time.

[0793] "Sales forecasting" refers to the use of machine learning models and other data to estimate future sales.

[0794] A "dashboard" is an interface designed for data visualization and analysis, and displays multiple pieces of information in an integrated manner.

[0795] "Inventory management proposal" is a method of proposing appropriate inventory levels and product placement based on past data and the latest information.

[0796] "PDF" is an abbreviation for Portable Document Format, a standard format for electronically storing and sharing documents.

[0797] This invention combines a system that collects past sales data, weather data, and economic indicator data and uses machine learning models to forecast sales, with an emotion engine that recognizes user emotions. This system is designed to improve sales efficiency and protect the environment, and also enables flexible responses based on user emotions.

[0798] The server comprises the following means:

[0799] 1. A means of collecting historical sales data, weather data, and economic indicator data

[0800] 2. Preprocessing the collected data, imputing missing values, and normalizing them

[0801] 3. A means to train machine learning models on preprocessed data

[0802] 4. Real-time recognition of user emotions and the use of emotional data

[0803] 5. A means of generating a sales forecast by inputting the latest data and sentiment data in response to a sales forecast request from a user.

[0804] 6. A means to display the generated sales forecast results on the user's device and visually display them on a dashboard

[0805] This will enable flexible sales forecasting and inventory management based on user sentiment, as well as customer service in physical stores.

[0806] Hardware and software used

[0807] Hardware

[0808] 1. Server: Performs key processing such as data collection, preprocessing, model training, and prediction generation.

[0809] 2. Smart glasses: Devices that detect the user's facial expressions and vocal tone to analyze emotions.

[0810] software

[0811] 1. OpenCV: An open-source library used for face detection and image processing.

[0812] 2. TensorFlow: A platform used to analyze facial expressions and recognize emotions using deep learning models.

[0813] 3. Scikit-Learn: A library used to perform sales forecasting using random forest regression models.

[0814] Example of a system

[0815] 1. Data Collection:

[0816] The server retrieves sales data for the past five years from an internal database, and also retrieves weather data (temperature, precipitation, etc.) for the past five years from meteorological agencies via API, as well as economic indicator data (GDP growth rate, unemployment rate).

[0817] 2. Data Preprocessing:

[0818] The server normalizes the collected data by imputing missing values, removing noise, and scaling. Missing sales data is imputed based on similar patterns in the past.

[0819] 3. Model training:

[0820] Using the pre-processed data, the server trains a random forest regression model to learn about past sales patterns and the effects of weather and economic conditions.

[0821] 4. Emotion recognition:

[0822] The smart glasses analyze the user's facial expressions and vocal tone in real time and use TensorFlow to obtain emotional data, which can determine whether the user is stressed or relaxed, for example.

[0823] 5. Sales forecast generation:

[0824] When a user requests "Sales forecast for October 2023," the server retrieves the latest weather forecast, economic indicator data, and sentiment data, and inputs these into the model to generate a sales forecast. If the server determines that the user is stressed, it will carefully provide feedback on the forecast and present the results.

[0825] 6. Results display:

[0826] The server sends the forecast results and inventory management suggestions to the user's device, which then visually displays them on a dashboard. The display content is also adjusted and the interface is customized based on the user's emotions.

[0827] Prompt Sentence Examples

[0828] "Local time in Tokyo on October 15, 2023, the weather is rainy, the temperature is 15 degrees, the GDP growth rate is 2.5%, the unemployment rate is 3%, and customer sentiment is stressed. Based on this data, generate a sales forecast."

[0829] This allows the present invention to leverage user sentiment and external data to optimize sales and customer experience in physical stores.

[0830] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0831] Step 1: Data collection

[0832] The server periodically collects historical sales data, weather data, and economic indicator data. The sales data is obtained from an internal company database, the weather data is obtained from a weather agency's API, and the economic indicator data is obtained from an economic database. The input is data from each data source, and the output is the collected data set.

[0833] Step 2: Data Preprocessing

[0834] The server preprocesses the collected data. Specifically, it imputes missing values ​​in the data, removes noise, and scales and normalizes it. For example, if there are gaps in sales data, it imputes them using similar data from the past. The input is the data collected in step 1, and the output is the preprocessed data.

[0835] Step 3: Model training

[0836] The server uses the preprocessed data to train a machine learning model. Training techniques such as random forest regression and LSTM (long short-term memory) networks are used. The model learns sales patterns based on past data. The input is the preprocessed data from step 2, and the output is the trained machine learning model.

[0837] Step 4: Emotion Recognition

[0838] The device analyzes the user's facial expressions and voice tone, and uses an emotion engine to recognize the user's emotions. The input is the user's voice and facial expression data acquired from the device, and the output is the analyzed emotion data.

[0839] Step 5: Generate a sales forecast

[0840] When a user sends a sales forecast request from their device, the server retrieves the latest weather and economic indicator data and inputs this data and sentiment data into the trained model. The server then adjusts the forecast and generates a sales forecast. The input is the user's request, the latest data, and sentiment data, and the output is the sales forecast result.

[0841] Step 6: View the results

[0842] The server generates forecast results and inventory management proposals and sends them to the user's terminal, where they are visually displayed as a dashboard. The input is the sales forecast results, and the output is the visually displayed data.

[0843] This process flow allows the present invention to respond to user emotions while providing more accurate sales forecasts and inventory management.

[0844] 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.

[0845] 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.

[0846] 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.

[0847] [Third embodiment]

[0848] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0849] 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.

[0850] 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).

[0851] 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.

[0852] 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.

[0853] 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).

[0854] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0855] 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.

[0856] 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.

[0857] 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.

[0858] 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.

[0859] 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."

[0860] The present invention is a system that collects past sales data, weather data, and economic indicator data, and uses machine learning models to forecast sales based on this data. This system is designed to improve sales efficiency and protect the environment, and an embodiment of the system is described below.

[0861] System Overview

[0862] This system consists of a server, a user's device, and a network infrastructure that connects them. The server has multiple functions, such as data collection, preprocessing, model training, prediction generation, and result display. Meanwhile, users input requests through their device and check the prediction results.

[0863] Program processing flow

[0864] 1. Data Collection

[0865] The server periodically collects sales data, weather data, and economic indicator data. The sales data is obtained from an internal database, the weather data is obtained from a weather agency's API, and the economic indicator data is obtained from an economic database.

[0866] 2. Data Preprocessing

[0867] The server preprocesses the collected data by imputing missing values, removing noise, scaling, and normalizing it. For example, for periods where sales data is missing, it uses similar data from the past to impute the missing data.

[0868] 3. Model Training

[0869] The server uses the preprocessed data to train a machine learning model, which uses a selected algorithm (e.g., random forest regression or LSTM) for sales forecasting. The model learns sales patterns based on past data.

[0870] 4. Prediction generation

[0871] When a user sends a sales forecast request from their device, the server retrieves the latest weather data and economic trend data from the API and uses this data to generate a sales forecast in real time.

[0872] 5. Display results

[0873] The server generates the prediction results and sends them to the user's device, where they can view the results visually on a dashboard and optionally download the report in PDF format.

[0874] Specific examples

[0875] Data collection

[0876] The server retrieves sales data for the past five years from an internal database, weather data (temperature, precipitation, etc.) for the past five years from the Japan Meteorological Agency via an API, and economic indicator data (e.g., GDP growth rate, unemployment rate) from the government and reliable economic forums.

[0877] Data Preprocessing

[0878] The server fills in missing data from the collected sales data using past approximate data. It also scales the sales data and weather data to a range of 0 to 1 and converts them into a format suitable for the model.

[0879] Model learning

[0880] The server trains a random forest regression model on the preprocessed data, which is used to learn about past sales patterns and the effects of weather and economic conditions and to predict future sales.

[0881] Forecast Generation

[0882] When a user requests "Sales forecast for October 2023," the server retrieves the latest weather forecast and economic indicators and inputs them into the trained model, which then outputs detailed sales forecasts and supply and demand plans for each region.

[0883] Results display

[0884] The server organizes the generated sales forecast results and displays them on the user's dashboard, where users can view the data in graphs and tables and download the reports in PDF format if required.

[0885] This frees sales representatives from manual work, enabling more efficient sales activities and achieving sustainable inventory management across the entire company.

[0886] The processing flow will be explained below.

[0887] Step 1:

[0888] The server retrieves past sales data from an internal database, which stores information such as sales date, product ID, sales quantity, price, and area.

[0889] Step 2:

[0890] The server obtains weather data from the weather agency's API, including past temperature, precipitation, wind speed, and air pressure.

[0891] Step 3:

[0892] The server retrieves economic indicator data from economic databases or trusted government APIs, such as GDP growth, unemployment rate, and consumer confidence index.

[0893] Step 4:

[0894] The server aggregates the collected sales, weather, and economic data, making the data set available in a consistent manner.

[0895] Step 5:

[0896] The server preprocesses the integrated data, imputing missing values, removing noise, and scaling the data. For example, if there is missing sales data, it is imputed using similar data from the past.

[0897] Step 6:

[0898] The server normalizes the data by scaling each variable to a range between 0 and 1. This step is important for improving the accuracy of training machine learning models.

[0899] Step 7:

[0900] The server uses the preprocessed data to train machine learning models, for example, using random forest regression or LSTM (long short-term memory) networks.

[0901] Step 8:

[0902] The server uses validation data to optimize model parameters during training and prevent overfitting or underfitting, a process that evaluates how well the model performs on real-world data.

[0903] Step 9:

[0904] The server compares the predictions with actual sales data to evaluate the accuracy of the model and tunes the model as needed, a step that is important for continuously improving the model's accuracy.

[0905] Step 10:

[0906] A user enters a request into a terminal, such as "Sales forecast for October 2023," which starts the forecast generation process.

[0907] Step 11:

[0908] The server retrieves the latest weather and economic indicator data in real time and inputs this into the trained model.

[0909] Step 12:

[0910] The server generates a sales forecast, which includes projected sales figures, peak demand times, and geographical details.

[0911] Step 13:

[0912] The server also generates inventory management recommendations based on the generated forecasts, helping to prevent overstocks and out-of-stock situations.

[0913] Step 14:

[0914] The server sends the prediction results and inventory management suggestions to the user's terminal.

[0915] Step 15:

[0916] The device visually displays the received prediction results on a dashboard, allowing users to view the results in graph and table format.

[0917] Step 16:

[0918] Users can check the forecast results on the dashboard and download the report in PDF format as needed, which allows them to develop specific sales strategies and inventory management plans.

[0919] As described above, the system collects data, preprocesses it, trains models, generates predictions, and displays the results step by step, supporting efficient sales activities and sustainable inventory management.

[0920] Example 1

[0921] 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."

[0922] Conventional sales forecasting systems often make predictions based solely on past sales data, without taking into account multiple factors such as weather data and economic indicators, resulting in low accuracy. Another issue is that few systems provide sales forecast results in real time, and the effects of missing data and noise cannot be fully mitigated. Furthermore, there is a lack of functionality that allows users to easily check sales forecast results and provides them in a format that can be used as reference material. These issues make it difficult to improve sales efficiency and protect the environment.

[0923] 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.

[0924] In this invention, the server includes: means for periodically collecting past sales data, weather data, and economic indicator data from an internal company database or an external API; means for preprocessing the collected data, filling in missing values ​​with past approximate data, removing noise, scaling, and normalizing the data; means for training a machine learning model based on the preprocessed data to learn past sales patterns and environmental factors; means for inputting the latest weather data and economic indicator data in response to a sales forecast request from a user and generating a sales forecast in real time; and means for transmitting the generated sales forecast results to the user's device, visually displaying them on a dashboard, and making them available for download as a PDF report if necessary. This improves the accuracy of sales forecasts and enables real-time information provision and visual display of results. As a result, sales efficiency and sustainable inventory management can be achieved.

[0925] An "in-house database" is a system installed within a company that stores and manages information, and holds business data such as sales data and customer information.

[0926] An "external organization's API" is an API (Application Programming Interface) provided by an external organization or service, and is a means of programmatically obtaining information such as weather data and economic indicator data.

[0927] "Sales data" refers to numerical information about sales recorded by a company during a specific period, such as the sales quantity and sales amount of a product.

[0928] "Weather data" is numerical information about the weather in a specific region or period, such as temperature, precipitation, humidity, and wind speed.

[0929] "Economic indicator data" refers to various statistical data that show the state of the economy, such as GDP growth rate, unemployment rate, and consumer price index.

[0930] "Preprocessing" refers to the preparation of data prior to data analysis or machine learning, and specifically includes processes such as filling in missing values, removing noise, scaling, and normalization.

[0931] "Missing values" refers to a state in which some data is missing, including blanks and NULL values ​​in a dataset.

[0932] "Noise" refers to unnecessary disturbances and errors contained in data that can reduce the accuracy of analyses and models.

[0933] "Scaling" is a conversion technique used to fit data values ​​into a range, usually normalizing the data range to 0 to 1 using minimum and maximum values.

[0934] "Normalization" refers to a transformation process to make the distribution of data uniform, particularly by confining the data to a specific range through scaling.

[0935] A "machine learning model" is a set of algorithms that learn patterns and rules from data and make predictions or classifications. Examples include random forests and LSTM.

[0936] "Training" is the process by which a machine learning model uses data to learn optimal parameters.

[0937] A "sales forecast request" is a request sent by a user to a server to forecast future sales based on a specific period and conditions.

[0938] "Real-time" refers to a state in which data is processed and displayed immediately at the moment it is generated or acquired.

[0939] A "dashboard" is a visual interface that allows users to see important information at a glance, displaying data in the form of graphs and tables.

[0940] "PDF format" stands for Portable Document Format, a file format for storing and sharing information in a fixed-layout format.

[0941] A "report" refers to the analysis results or report on specific data, and is often generated in PDF format or similar.

[0942] This invention is a system that collects past sales data, weather data, and economic indicator data from in-house databases and APIs of external institutions, and uses this data to perform sales forecasts using machine learning models. This system consists of a server, user terminals, and a network infrastructure that connects these.

[0943] Hardware and Software Configuration

[0944] server

[0945] The server is a computer system equipped with a high-performance processor and large memory capacity, and is installed with software for database management systems, API handlers, and machine learning models, including pandas, scikit-learn, and TensorFlow for data analysis.

[0946] Terminal

[0947] Users access the system through a web browser on devices such as PCs and tablets. The device must have an internet connection and a browser (e.g., Google Chrome or Mozilla Firefox) to view dashboards and reports.

[0948] Data collection

[0949] The server periodically connects to the company's internal database and retrieves sales data for the past five years, which is retrieved using SQL queries.

[0950] The server retrieves weather data for the past five years through the API of an external weather agency. This API request uses the HTTP GET method to retrieve data such as temperature and precipitation for a specified period.

[0951] The server uses APIs provided by governments and economic forums to collect economic indicator data (e.g., GDP growth rate, unemployment rate) for the past five years.

[0952] Data Preprocessing

[0953] The server uses the pandas library to fill in missing values ​​in the acquired sales data with similar historical data, and applies appropriate filtering techniques to remove noise.

[0954] Next, we use scikit-learn's MinMaxScaler to scale and normalize the sales and weather data to the range 0 to 1, so that the data can be fed into the machine learning model in a unified format.

[0955] Model learning

[0956] The server trains a machine learning model based on the preprocessed data. Algorithms such as random forest regression models and LSTM are used, and the model is run using scikit-learn and TensorFlow. During the training process, the dataset is split into training data and validation data, and cross-validation is applied to ensure the accuracy of the model.

[0957] Forecast Generation

[0958] When a user submits a sales forecast request from their device, they enter a prompt, such as "Sales forecast for October 2023." In response, the server again retrieves the latest weather and economic data and inputs it into the trained model.

[0959] The model generates results by predicting future sales in real time based on new data.

[0960] Results display

[0961] The server organizes the generated prediction results and sends them to the user's device, where they are visually displayed in graphs and tables on a dashboard for the user to review.

[0962] If desired, users can download the report in PDF format, a feature that uses libraries such as matplotlib and ReportLab.

[0963] Specific examples

[0964] Data collection

[0965] The server retrieves sales data for the past five years from a company database, collects weather data from the Japan Meteorological Agency's API, and obtains economic indicator data from the government's economic forum.

[0966] Data Preprocessing

[0967] The server fills in missing sales data with similar data from the past and scales the weather data to normalize it to the range 0 to 1.

[0968] Model learning

[0969] The server trains a random forest regression model to learn patterns from the data.

[0970] Forecast Generation

[0971] When a user requests "Sales forecast for October 2023," the server inputs the latest data into the model and generates a sales forecast.

[0972] Results display

[0973] The server displays the prediction results on a dashboard and allows users to download the report in PDF format.

[0974] Prompt Sentence Examples

[0975] "Show me sales forecast for October 2023."

[0976] "Get weather data for the past five years."

[0977] "Please update your economic indicator data."

[0978] This system frees sales representatives from manual work, enabling more efficient sales activities and achieving sustainable inventory management across the entire company.

[0979] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0980] Step 1:

[0981] The server connects to the company's internal database and retrieves sales data for the past five years. This is done using an SQL query. Specifically, the server executes a query such as "SELECT FROM sales WHERE date BETWEEN '2018-01-01' AND '2022-12-31'" to extract the required data. The input is the connection information and the sales data contained in the database, and the output is the sales data for the specified period.

[0982] Step 2:

[0983] The server sends an HTTP GET request to the weather agency's API to retrieve weather data for the past five years. Specifically, it sends a request such as "GET http: / / api.weather.com / data?startDate=2018-01-01&endDate=2022-12-31" to retrieve temperature and precipitation data. The input is the API endpoint information, and the output is the weather data returned by the API.

[0984] Step 3:

[0985] The server accesses the government's economic data API to obtain economic indicator data (e.g., GDP growth rate and unemployment rate) for the past five years. Specifically, it executes "GET http: / / api.economicdata.com / indicators?startDate=2018-01-01&endDate=2022-12-31" to obtain the indicator data. The input is the API endpoint information, and the output is the obtained economic indicator data.

[0986] Step 4:

[0987] The server preprocesses the collected sales data, weather data, and economic indicator data. Specifically, it creates a data frame using the pandas library and imputes missing values ​​using methods such as linear imputation. The input is the collected raw data, and the output is a data frame with missing values ​​imputed.

[0988] Step 5:

[0989] The server scales and normalizes the preprocessed data, specifically using scikit-learn's MinMaxScaler to convert values ​​to the range 0 to 1. The input is the interpolated data frame, and the output is the scaled data.

[0990] Step 6:

[0991] The server uses the scaled data to train a machine learning model, for example, a random forest regression model using scikit-learn's RandomForestRegressor. The input is the preprocessed and scaled data, and the output is a trained machine learning model.

[0992] Step 7:

[0993] The user sends a sales forecast request from the terminal. For example, they input a prompt such as "Sales forecast for October 2023." The input is the user's request, and the output is the request data to the server.

[0994] Step 8:

[0995] The server retrieves the latest weather data and economic indicator data from the API again and inputs it into the trained model, generating a sales forecast in real time. The input is the latest weather data and economic indicator data, and the output is sales forecast data.

[0996] Step 9:

[0997] The server organizes the generated sales forecast results and sends them to the user's device. The data is displayed visually on a dashboard. The input is the sales forecast data, and the output is the graphs and tables displayed on the dashboard.

[0998] Step 10:

[0999] Users can check the forecast results through a dashboard and download them in PDF format if necessary. The server generates reports using libraries such as matplotlib and ReportLab. The input is the user's download request and sales forecast data, and the output is a PDF report.

[1000] (Application example 1)

[1001] 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."

[1002] In traditional manufacturing, demand forecasting and inventory management are often performed manually, resulting in problems such as human error and reduced efficiency. It is also difficult to flexibly adjust manufacturing processes based on real-time data from the field, making it difficult to maintain a balance between supply and demand. Furthermore, it is difficult to make accurate predictions that take environmental fluctuations (weather and economic conditions) into account, leading to inventory surpluses and shortages. This calls for a method to improve production efficiency, minimize resource waste, and establish a sustainable production system.

[1003] 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.

[1004] In this invention, the server includes: means for collecting past sales data, weather data, and economic index data; means for preprocessing the collected data, imputing missing values, and normalizing the data; means for training a machine learning model based on the preprocessed data; means for inputting the latest data and generating a sales forecast in response to a sales forecast request from a user; means for displaying the generated sales forecast results on the user's terminal and visually showing them on a dashboard; and means for instructing a robot to adjust the manufacturing process based on the generated forecast results. This allows the factory's manufacturing process to be automatically adjusted in real time based on supply and demand forecasts, not only improving production efficiency but also reducing human error and enabling sustainable inventory management.

[1005] "Past sales data" refers to sales performance information recorded by companies and stores to date, specifically information such as product sales quantity, sales amount, and sales period.

[1006] "Weather data" refers to meteorological information such as temperature, precipitation, and wind speed obtained from meteorological agencies and used in forecasting models.

[1007] "Economic indicator data" refers to data that shows economic trends, and includes GDP growth rate, unemployment rate, consumer price index, etc.

[1008] "Means of collection" refers to the methods and technologies used to obtain the necessary information from various data sources.

[1009] "Preprocessing means" refers to methods and techniques for converting acquired data into an analyzable format, and specifically includes imputing missing values ​​and normalizing data.

[1010] "Methods for imputing missing values" are methods for completing incomplete data by using similar data from the past or average values, etc.

[1011] "Normalization" is a method of scaling data values ​​into a certain range so that machine learning algorithms can learn efficiently.

[1012] A "means for training a machine learning model" is a method for training a machine learning algorithm using preprocessed data to build a predictive model.

[1013] The "means for generating sales forecasts" is a method for inputting the latest data into a trained machine learning model to predict future sales.

[1014] A "user's terminal" is a device such as a PC or smartphone that the user uses to check the prediction results.

[1015] A "dashboard" refers to an interface that allows users to visually check prediction results, displaying information in the form of graphs and tables.

[1016] "Means for issuing instructions to robots to adjust the manufacturing process based on the generated prediction results" refers to methods and technologies for issuing specific operating instructions to manufacturing robots based on predicted supply and demand data, automatically adjusting the production line.

[1017] The present invention is a system that collects past sales data, weather data, and economic indicator data, and uses a machine learning model based on this data to forecast sales, and this system has been applied to factory robots. The system is configured as follows.

[1018] System Overview

[1019] This system consists of a server, user terminals, factory robots, and a network infrastructure that connects them. The server has multiple functions, including data collection, preprocessing, model learning, prediction generation, and result display. Meanwhile, users input requests through terminals or consoles and check prediction results. Furthermore, the factory robots automatically adjust the manufacturing process based on this prediction data.

[1020] Hardware and software used

[1021] Hardware:

[1022] Factory robots (e.g. Fanuc, ABB, KUKA, etc.)

[1023] Servers (cloud or on-site data center)

[1024] Sensors (monitor temperature, humidity, vibration, etc.)

[1025] software:

[1026] Machine learning frameworks (e.g. TensorFlow, PyTorch)

[1027] Database (e.g. MySQL, PostgreSQL)

[1028] API services (e.g. OpenWeatherMap, economic data API)

[1029] Robot control software (e.g. ROS - Robot Operating System)

[1030] Data processing and calculation

[1031] 1. The server retrieves past sales data from the company's internal database, collects weather data from the weather agency's API, and collects economic indicator data from the economic database.

[1032] 2. The server preprocesses the collected data. Specifically, it imputes missing values, removes noise, and scales and normalizes the data. For example, if sales data is missing for a period, it is imputed using similar data from the past.

[1033] 3. The server uses the preprocessed data to train a machine learning model. This model uses a selected algorithm (e.g., random forest regression or LSTM) for sales forecasting. The model learns sales patterns based on past data.

[1034] 4. When a user sends a sales forecast request from their device, the server retrieves the latest weather data and economic trend data from the API and uses this data to generate a sales forecast in real time.

[1035] 5. The server generates the prediction results and sends them to the user's device. The user can view the results visually on a dashboard and optionally download a report in PDF format.

[1036] 6. Based on the generated prediction results, the server sends instructions to the factory robot to adjust the manufacturing process, allowing the factory robot to automatically optimize the manufacturing process based on the supply and demand balance.

[1037] Specific examples

[1038] Data collection:

[1039] The server retrieves the production and inventory data for the past two years from the factory database, and also uses APIs to retrieve weather data (e.g., temperature, humidity) for the past two years, and collects economic data (e.g., GDP growth rate, production index).

[1040] Data preprocessing:

[1041] The server fills in missing data using past approximate data, scales sales data and weather data to a range of 0 to 1, and converts them into a format suitable for the model.

[1042] Model training:

[1043] The server trains a random forest regression model on the preprocessed data, which is used to learn about past sales patterns and the effects of weather and economic conditions and to predict future sales.

[1044] Forecast generation:

[1045] When a user requests "Sales forecast for October 2023," the server retrieves the latest weather forecast and economic indicators and inputs them into the trained model, which then outputs detailed sales forecasts and supply and demand plans for each region.

[1046] Results display and automatic adjustment:

[1047] The server then organizes the generated sales forecasts and displays them on the user's dashboard. Based on the forecasts, the server sends instructions to factory robots to adjust the manufacturing process, for example, adjusting the production speed of a specific product line to prevent overstocks or shortages.

[1048] Example prompts to input to the generative AI model

[1049] Input historical production data, weather data, and economic indicator data to generate a production demand forecast for the next month.

[1050] Using this system, supply and demand forecasts and production process optimization at manufacturing sites can be performed in real time, improving production efficiency and optimizing inventory management.

[1051] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1052] Step 1: Data collection

[1053] The server retrieves past sales data from the company's internal database, and uses APIs to collect weather data from meteorological agencies and economic indicator data from economic databases. As a result, past sales data, weather data, and economic indicator data are aggregated on the server as input data. Specifically, this includes sales quantity, sales amount, temperature, precipitation, GDP growth rate, unemployment rate, etc.

[1054] Step 2: Data Preprocessing

[1055] The server then fills in missing values ​​in the collected data, removes noise, and scales and normalizes the data. Specifically, it fills in missing sales data using similar historical data, and does the same for weather data and economic indicator data. The result is a preprocessed dataset that can be used by machine learning models.

[1056] Step 3: Train the machine learning model

[1057] The server uses the preprocessed data to train a machine learning model for sales forecasting. Specifically, it uses algorithms such as random forest regression and LSTM to learn sales patterns and correlations with various metrics, resulting in a trained machine learning model.

[1058] Step 4: Receive a prediction request

[1059] The user sends a sales forecast request from the terminal. For example, they enter a prompt sentence such as "Please generate sales forecast for October 2023." This becomes the input data to the server.

[1060] Step 5: Generate forecasts

[1061] The server receives the request from the user and retrieves the latest weather and economic indicator data from the API again. The retrieved data is input into the trained model to generate a sales forecast. This operation outputs real-time sales forecast data.

[1062] Step 6: View the results

[1063] The server sends the generated sales forecast results to the user's device, where the user can visually check the forecast results on the dashboard of their device. For example, the sales forecast data can be displayed in graph or table format.

[1064] Step 7: Adjust the manufacturing process based on the results

[1065] Based on the generated predictions, the server sends instructions to the factory robots to adjust the manufacturing process. The factory robots then use the instructions received from the server to optimize the manufacturing process in real time, for example, by adjusting the production speed of a particular product line or changing the amount of materials required.

[1066] Through these processing steps, the system automates everything from sales forecasts to real-time adjustments to manufacturing processes, enabling efficient production management.

[1067] 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.

[1068] This invention combines a system that collects past sales data, weather data, and economic indicator data and uses machine learning models to forecast sales, with an emotion engine that recognizes user emotions. This system is designed to improve sales efficiency and protect the environment, and also enables flexible responses based on user emotions.

[1069] System Overview

[1070] This system consists of a server, a user's device, an emotion engine, and a network infrastructure that connects these components. The server has multiple functions, including data collection, preprocessing, model training, prediction generation, result display, and emotion recognition. Meanwhile, users input requests through their device and check prediction results.

[1071] Program processing flow

[1072] 1. Data Collection

[1073] The server periodically collects historical sales data, weather data, and economic indicator data. The sales data is obtained from an internal database, the weather data is obtained from a weather agency's API, and the economic indicator data is obtained from an economic database.

[1074] 2. Data Preprocessing

[1075] The server preprocesses the collected data by filling in missing values, removing noise, scaling, and normalizing it. For example, if there are gaps in sales data, it fills them in using similar data from the past.

[1076] 3. Model Training

[1077] The server uses the preprocessed data to train a machine learning model using techniques such as random forest regression and LSTM (long short-term memory) networks. The model learns sales patterns based on past data.

[1078] 4. Emotion recognition

[1079] The device analyzes the user's facial expressions and tone of voice, and uses an emotion engine to recognize the user's emotions. The emotion data is sent to the server in real time.

[1080] 5. Prediction Generation

[1081] When a user sends a sales forecast request from their device, the server retrieves the latest weather and economic indicator data, inputs this data into the trained model, and adjusts the forecast taking into account emotion data from the emotion engine.

[1082] 6. Results display

[1083] The server generates prediction results and inventory management suggestions and sends them to the user's device. The user can view the results visually on a dashboard, and emotion data is used to dynamically adjust the display content and interface.

[1084] Specific examples

[1085] Data collection

[1086] The server retrieves sales data for the past five years from an internal database, and also retrieves weather data (temperature, precipitation, etc.) for the past five years from meteorological agencies via API, as well as economic indicator data (GDP growth rate, unemployment rate).

[1087] Data Preprocessing

[1088] The server normalizes the collected data by imputing missing values, removing noise, and scaling. Missing sales data is interpolated based on similar patterns in the past.

[1089] Model learning

[1090] The server trains a random forest regression model on the pre-processed data to learn about past sales patterns and the effects of weather and economic conditions.

[1091] emotion recognition

[1092] The device analyzes the user's facial expressions and voice tone in real time and obtains emotional data through an emotion engine, which can determine whether the user is, for example, stressed or relaxed.

[1093] Forecast Generation

[1094] When a user requests "Sales forecast for October 2023," the server retrieves the latest weather forecast, economic indicator data, and sentiment data, and inputs these into the model to generate a sales forecast. If the server determines that the user is stressed, it will carefully provide feedback on the forecast and present the results.

[1095] Results display

[1096] The server sends the forecast results and inventory management suggestions to the user's device, which displays them in graph and table format on a dashboard. The user can check the results and download the report in PDF format if necessary. The display content is also adjusted and the interface is customized based on the user's sentiment.

[1097] In this way, the present invention can provide more accurate sales forecasts and inventory management while responding to user emotions.

[1098] The processing flow will be explained below.

[1099] Step 1:

[1100] The server retrieves past sales data from an internal database, which stores information such as sales date, product ID, sales quantity, price, and area.

[1101] Step 2:

[1102] The server obtains weather data from the weather agency's API, specifically, past temperature, precipitation, wind speed, air pressure, etc.

[1103] Step 3:

[1104] The server retrieves economic indicator data from economic databases or trusted government APIs, such as GDP growth, unemployment rate, and consumer confidence index.

[1105] Step 4:

[1106] The server aggregates the collected sales data, weather data, and economic indicator data, allowing for a consistent analysis of the data set.

[1107] Step 5:

[1108] The server preprocesses the integrated data by imputing missing values, removing noise, and scaling it. For example, if there is missing sales data, it fills in the missing data using similar data from the past.

[1109] Step 6:

[1110] The server uses the preprocessed data as training data for machine learning models, such as random forest regression and LSTM (long short-term memory) networks, to train the models.

[1111] Step 7:

[1112] The server uses validation data to optimize the parameters of the trained model and prevent overfitting and underfitting.

[1113] Step 8:

[1114] The device analyzes the user's facial expressions and voice tone in real time and recognizes their emotions using an emotion engine. The recognized emotion data is then sent to the server.

[1115] Step 9:

[1116] When a user requests "Sales forecast for October 2023" from their device, the server retrieves the latest weather data, economic indicator data, and sentiment data.

[1117] Step 10:

[1118] The server then inputs this data into a trained model to generate sales forecasts, with the sentiment data being used to fine-tune the results.

[1119] Step 11:

[1120] The server organizes the generated sales forecast results and simultaneously creates inventory management proposals, thereby preventing excess inventory and out-of-stock situations.

[1121] Step 12:

[1122] The server transmits the sales forecast results and inventory management proposals to the user's terminal.

[1123] Step 13:

[1124] The terminal displays the received sales forecast results and inventory management suggestions on a dashboard, including graphs and tables.

[1125] Step 14:

[1126] Users can check the forecast results on the dashboard and download the report in PDF format if necessary.

[1127] Step 15:

[1128] The device dynamically adjusts the display content and interface based on the user's emotions. For example, if the user is feeling stressed, the device will change the display to a simpler display and a more visually soothing color scheme.

[1129] As described above, the system performs data collection, preprocessing, model training, emotion recognition, prediction generation, and result display at each step, supporting efficient sales activities and sustainable inventory management while also improving the user experience.

[1130] Example 2

[1131] 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."

[1132] Conventional sales forecasting systems typically collect past sales data, weather data, economic indicator data, etc., and use machine learning to forecast sales. However, they were unable to take into account individual factors such as the user's emotional state. This meant that they were unable to respond flexibly to user emotions, making it difficult to improve forecast accuracy and user experience. Furthermore, there were limited ways to visually confirm forecast results and present them in a format that was easy for users to understand.

[1133] 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.

[1134] In this invention, the server includes: means for collecting past sales data, weather data, and economic indicator data; means for preprocessing the collected data, imputing missing values, and normalizing the data; means for training a machine learning model based on the preprocessed data; means for recognizing user emotion data and transmitting it to the server; means for inputting the latest data and emotion data and generating a sales forecast in response to a sales forecast request from the user; and means for displaying the generated sales forecast results on the user's terminal and visually presenting them on a dashboard. This enables sales forecasts that take the user's emotional state into account, improving forecast accuracy and the user experience. Furthermore, visually displaying the forecast results makes it possible to provide a system that is easy for users to understand and use.

[1135] "Past sales data" refers to records of past sales and purchases of goods conducted by a company, and includes data such as sales quantities, amounts, and dates.

[1136] "Weather data" refers to data including meteorological information such as temperature, precipitation, wind speed, and humidity obtained from meteorological agencies.

[1137] "Economic indicator data" is statistical data used to understand the state of the economy, such as gross domestic product (GDP), unemployment rate, and inflation rate.

[1138] "Preprocessing" refers to the processing performed on collected raw data, and involves data cleansing and normalization, such as filling in missing values, removing noise, and scaling.

[1139] A "machine learning model" is an algorithm or computational model that learns patterns and regularities from given data and makes predictions and judgments about future data.

[1140] "Emotion data" is data that indicates the emotional state of the user analyzed from facial expressions, tone of voice, and the like.

[1141] A "server" is a central computer that performs data collection, preprocessing, model training, prediction generation, and display of results.

[1142] A "terminal" is a computer or mobile device used by a user to enter data and confirm results.

[1143] A "dashboard" is an interface that allows users to visually check information such as sales forecast results and inventory management proposals.

[1144] This invention is a system that collects and analyzes past sales data, weather data, economic indicator data, and even user emotion data, and uses machine learning models to forecast sales. This system is composed of a server, a terminal, an emotion recognition engine, and a network infrastructure that links these components. Below, we will explain in detail how each component works.

[1145] Data collection

[1146] The server periodically collects past sales data, weather data, and economic indicator data. Sales data is obtained from an internal database, weather data is obtained using the weather agency's API, and economic indicator data is obtained from an economic database. Specifically, sales data is extracted from the database by executing SQL queries using Python. Weather data is obtained in JSON format from the weather agency's API using an HTTP request, for example, and economic indicator data is collected from the IMF and OECD APIs.

[1147] Data Preprocessing

[1148] The server preprocesses the collected data. This preprocessing includes imputing missing values, removing noise, scaling, and normalizing the data. For example, it uses Python's pandas library to impute missing values ​​in the data frame with the mean value and scales the data using scikit-learn's StandardScaler. This cleans the data so that the machine learning model can operate optimally.

[1149] Model learning

[1150] The server uses the preprocessed data to train a machine learning model. The main techniques used are random forest regression and LSTM (long short-term memory). For example, a model is created using scikit-learn's RandomForestRegressor and trained using the training data. The server also evaluates the accuracy of the model and adjusts hyperparameters using cross-validation if necessary.

[1151] emotion recognition

[1152] The device captures the user's facial expressions and voice tone and analyzes them with an emotion recognition engine. For example, it uses OpenCV to analyze video from a webcam and recognize emotions from facial expressions. It also uses the Librosa library to extract voice features and classify emotions using a deep learning model. This emotion data is sent to a server in real time and incorporated into a prediction algorithm.

[1153] Forecast Generation

[1154] When a user requests a sales forecast through their device, the server retrieves the latest weather data, economic indicator data, and emotion data and inputs this into the trained model. For example, if a user requests "Sales forecast for October 2023," the server collects the data and generates a sales forecast. The user's emotion data is also taken into account and the forecast results are adjusted. For example, a more conservative forecast is provided to a user who is feeling stressed.

[1155] Results display

[1156] The server generates forecast results and inventory management proposals and sends them to the user's device. The device visually displays the results on a dashboard, allowing the user to intuitively understand the results. For example, graphs and tables are created using the Plotly library and provided to the user. The user can check the results and download the report in PDF format if necessary. The color and layout of the user interface are also dynamically adjusted based on the user's mood.

[1157] Example prompts to input to the generative AI model

[1158] "Please provide a sales forecast for October 2023. Please use the latest weather and economic indicator data. Please also take into account the user's emotional state and provide results."

[1159] In this way, the present invention enables sales forecasting that takes into account the emotional state of the user, thereby realizing a system that provides higher accuracy and a better user experience.

[1160] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1161] Step 1: Data collection

[1162] The server collects historical sales data, weather data, and economic indicator data.

[1163] Input: In-house databases, weather agency APIs, economic databases.

[1164] What it does: Uses Python to run SQL queries to extract sales data. Uses HTTP requests to retrieve weather data from meteorological agency APIs and economic indicators data from IMF and OECD APIs. Saves this data in CSV and JSON formats.

[1165] Output: Files with historical sales data, weather data, and economic indicator data.

[1166] Step 2: Data Preprocessing

[1167] The server pre-processes the collected data.

[1168] Inputs: Historical sales data, weather data, economic indicator data.

[1169] Specific operation: Uses the pandas library to impute missing values ​​in a data frame with the mean and remove outliers. Scales the data using scikit-learn's StandardScaler, which imputes missing values, removes noise, and scales the data.

[1170] Output: Preprocessed data.

[1171] Step 3: Model training

[1172] The server trains a machine learning model using the preprocessed data.

[1173] Input: Preprocessed data.

[1174] Specific operation: Create a model using scikit-learn's RandomForestRegressor, train the model using the training data, evaluate the model's accuracy using cross-validation, and adjust the hyperparameters as needed.

[1175] Output: A trained machine learning model.

[1176] Step 4: Emotion Recognition

[1177] The terminal recognizes the user's emotion data.

[1178] Input: User's facial expressions and vocal tone.

[1179] Specific operation: Analyzes video from a webcam using OpenCV and recognizes emotions from facial expressions. Extracts audio features using the Librosa library and classifies emotions using a deep learning model. Sends emotion data to the server in real time using WebSocket.

[1180] Output: Real-time sentiment data.

[1181] Step 5: Generate forecasts

[1182] The user submits a sales forecast request through a terminal.

[1183] Inputs: Sales forecast request from user, weather updates, economic indicators, and sentiment data.

[1184] How it works: The server receives the request and inputs the latest weather data, economic indicators, and sentiment data into the model. It then adjusts the prediction results to account for the user's emotional state and generates a sales forecast.

[1185] Output: Sales forecast results.

[1186] Step 6: View the results

[1187] The server generates the forecast results and inventory management proposals and sends them to the user's terminal.

[1188] Input: Sales forecast results, inventory management proposals.

[1189] Specific operation: The server generates forecast results and inventory management suggestions in JSON format and sends them to the terminal. The terminal uses the Plotly library to visually display them as graphs and tables on a dashboard. The UI color and layout are dynamically adjusted according to the user's mood. The user can also download the results in PDF format.

[1190] Output: Forecast results displayed in a dashboard, a report in PDF format.

[1191] (Application example 2)

[1192] 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."

[1193] Conventional sales forecasting systems make predictions based on past data and environmental data, but because they do not take user emotions into account, they often lack the accuracy of predictions and flexibility in customer service. Furthermore, in customer service situations in physical stores, there is no way to accurately grasp customer emotions, making it difficult to efficiently manage inventory and recommend appropriate products.

[1194] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting past sales data, weather data, and economic index data, means for preprocessing the collected data, imputing missing values, and normalizing the data, means for training a machine learning model based on the preprocessed data, means for recognizing user emotions in real time and using the emotion data, means for inputting the latest data and emotion data and generating a sales forecast in response to a sales forecast request from a user, and means for displaying the generated sales forecast results on the user's terminal and visually showing them on a dashboard. This enables flexible sales forecasting and inventory management based on user emotions, as well as customer service in physical stores.

[1195] "Historical sales data" refers to sales information recorded based on previous sales activity.

[1196] "Weather data" refers to numerical data or information that represents meteorological information, and is data that indicates the weather conditions at a specific date, time, and location.

[1197] "Economic indicator data" is data based on multiple economic indicators that show economic trends and conditions.

[1198] "Preprocessing" is the process of preparing collected data using methods such as filling in missing values, removing noise, and scaling.

[1199] A "machine learning model" is an algorithm that learns patterns and rules from large amounts of data and performs tasks such as prediction and classification.

[1200] "User emotion data" is information that indicates the user's emotional state in real time, obtained from the user's facial expressions, tone of voice, and the like.

[1201] An "emotion engine" is a technology or device that analyzes input data such as a user's facial expressions and voice to determine their emotional state.

[1202] A "sales forecast request" is a request made by a user to the system to forecast future sales.

[1203] "Latest data" refers to the most recent information or data available at the present time.

[1204] "Sales forecasting" refers to the use of machine learning models and other data to estimate future sales.

[1205] A "dashboard" is an interface designed for data visualization and analysis, and displays multiple pieces of information in an integrated manner.

[1206] "Inventory management proposal" is a method of proposing appropriate inventory levels and product placement based on past data and the latest information.

[1207] "PDF" is an abbreviation for Portable Document Format, a standard format for electronically storing and sharing documents.

[1208] This invention combines a system that collects past sales data, weather data, and economic indicator data and uses machine learning models to forecast sales, with an emotion engine that recognizes user emotions. This system is designed to improve sales efficiency and protect the environment, and also enables flexible responses based on user emotions.

[1209] The server comprises the following means:

[1210] 1. A means of collecting historical sales data, weather data, and economic indicator data

[1211] 2. Preprocessing the collected data, imputing missing values, and normalizing them

[1212] 3. A means to train machine learning models on preprocessed data

[1213] 4. Real-time recognition of user emotions and the use of emotional data

[1214] 5. A means of generating a sales forecast by inputting the latest data and sentiment data in response to a sales forecast request from a user.

[1215] 6. A means to display the generated sales forecast results on the user's device and visually display them on a dashboard

[1216] This will enable flexible sales forecasting and inventory management based on user sentiment, as well as customer service in physical stores.

[1217] Hardware and software used

[1218] Hardware

[1219] 1. Server: Performs key processing such as data collection, preprocessing, model training, and prediction generation.

[1220] 2. Smart glasses: Devices that detect the user's facial expressions and vocal tone to analyze emotions.

[1221] software

[1222] 1. OpenCV: An open-source library used for face detection and image processing.

[1223] 2. TensorFlow: A platform used to analyze facial expressions and recognize emotions using deep learning models.

[1224] 3. Scikit-Learn: A library used to perform sales forecasting using random forest regression models.

[1225] Example of a system

[1226] 1. Data Collection:

[1227] The server retrieves sales data for the past five years from an internal database, and also retrieves weather data (temperature, precipitation, etc.) for the past five years from meteorological agencies via API, as well as economic indicator data (GDP growth rate, unemployment rate).

[1228] 2. Data Preprocessing:

[1229] The server normalizes the collected data by imputing missing values, removing noise, and scaling. Missing sales data is imputed based on similar patterns in the past.

[1230] 3. Model training:

[1231] Using the pre-processed data, the server trains a random forest regression model to learn about past sales patterns and the effects of weather and economic conditions.

[1232] 4. Emotion recognition:

[1233] The smart glasses analyze the user's facial expressions and vocal tone in real time and use TensorFlow to obtain emotional data, which can determine whether the user is stressed or relaxed, for example.

[1234] 5. Sales forecast generation:

[1235] When a user requests "Sales forecast for October 2023," the server retrieves the latest weather forecast, economic indicator data, and sentiment data, and inputs these into the model to generate a sales forecast. If the server determines that the user is stressed, it will carefully provide feedback on the forecast and present the results.

[1236] 6. Results display:

[1237] The server sends the forecast results and inventory management suggestions to the user's device, which then visually displays them on a dashboard. The display content is also adjusted and the interface is customized based on the user's emotions.

[1238] Prompt Sentence Examples

[1239] "Local time in Tokyo on October 15, 2023, the weather is rainy, the temperature is 15 degrees, the GDP growth rate is 2.5%, the unemployment rate is 3%, and customer sentiment is stressed. Based on this data, generate a sales forecast."

[1240] This allows the present invention to leverage user sentiment and external data to optimize sales and customer experience in physical stores.

[1241] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1242] Step 1: Data collection

[1243] The server periodically collects historical sales data, weather data, and economic indicator data. The sales data is obtained from an internal company database, the weather data is obtained from a weather agency's API, and the economic indicator data is obtained from an economic database. The input is data from each data source, and the output is the collected data set.

[1244] Step 2: Data Preprocessing

[1245] The server preprocesses the collected data. Specifically, it imputes missing values ​​in the data, removes noise, and scales and normalizes it. For example, if there are gaps in sales data, it imputes them using similar data from the past. The input is the data collected in step 1, and the output is the preprocessed data.

[1246] Step 3: Model training

[1247] The server uses the preprocessed data to train a machine learning model. Training techniques such as random forest regression and LSTM (long short-term memory) networks are used. The model learns sales patterns based on past data. The input is the preprocessed data from step 2, and the output is the trained machine learning model.

[1248] Step 4: Emotion Recognition

[1249] The device analyzes the user's facial expressions and voice tone, and uses an emotion engine to recognize the user's emotions. The input is the user's voice and facial expression data acquired from the device, and the output is the analyzed emotion data.

[1250] Step 5: Generate a sales forecast

[1251] When a user sends a sales forecast request from their device, the server retrieves the latest weather and economic indicator data and inputs this data and sentiment data into the trained model. The server then adjusts the forecast and generates a sales forecast. The input is the user's request, the latest data, and sentiment data, and the output is the sales forecast result.

[1252] Step 6: View the results

[1253] The server generates forecast results and inventory management proposals and sends them to the user's terminal, where they are visually displayed as a dashboard. The input is the sales forecast results, and the output is the visually displayed data.

[1254] This process flow allows the present invention to respond to user emotions while providing more accurate sales forecasts and inventory management.

[1255] 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.

[1256] 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.

[1257] 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.

[1258] [Fourth embodiment]

[1259] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1260] 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.

[1261] 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).

[1262] 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.

[1263] 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.

[1264] 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).

[1265] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1266] 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.

[1267] 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.

[1268] 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.

[1269] 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.

[1270] 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.

[1271] 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."

[1272] The present invention is a system that collects past sales data, weather data, and economic indicator data, and uses machine learning models to forecast sales based on this data. This system is designed to improve sales efficiency and protect the environment, and an embodiment of the system is described below.

[1273] System Overview

[1274] This system consists of a server, a user's device, and a network infrastructure that connects them. The server has multiple functions, such as data collection, preprocessing, model training, prediction generation, and result display. Meanwhile, users input requests through their device and check the prediction results.

[1275] Program processing flow

[1276] 1. Data Collection

[1277] The server periodically collects sales data, weather data, and economic indicator data. The sales data is obtained from an internal database, the weather data is obtained from a weather agency's API, and the economic indicator data is obtained from an economic database.

[1278] 2. Data Preprocessing

[1279] The server preprocesses the collected data by imputing missing values, removing noise, scaling, and normalizing it. For example, for periods where sales data is missing, it uses similar data from the past to impute the missing data.

[1280] 3. Model Training

[1281] The server uses the preprocessed data to train a machine learning model, which uses a selected algorithm (e.g., random forest regression or LSTM) for sales forecasting. The model learns sales patterns based on past data.

[1282] 4. Prediction generation

[1283] When a user sends a sales forecast request from their device, the server retrieves the latest weather data and economic trend data from the API and uses this data to generate a sales forecast in real time.

[1284] 5. Display results

[1285] The server generates the prediction results and sends them to the user's device, where they can view the results visually on a dashboard and optionally download the report in PDF format.

[1286] Specific examples

[1287] Data collection

[1288] The server retrieves sales data for the past five years from an internal database, weather data (temperature, precipitation, etc.) for the past five years from the Japan Meteorological Agency via an API, and economic indicator data (e.g., GDP growth rate, unemployment rate) from the government and reliable economic forums.

[1289] Data Preprocessing

[1290] The server fills in missing data from the collected sales data using past approximate data. It also scales the sales data and weather data to a range of 0 to 1 and converts them into a format suitable for the model.

[1291] Model learning

[1292] The server trains a random forest regression model on the preprocessed data, which is used to learn about past sales patterns and the effects of weather and economic conditions and to predict future sales.

[1293] Forecast Generation

[1294] When a user requests "Sales forecast for October 2023," the server retrieves the latest weather forecast and economic indicators and inputs them into the trained model, which then outputs detailed sales forecasts and supply and demand plans for each region.

[1295] Results display

[1296] The server organizes the generated sales forecast results and displays them on the user's dashboard, where users can view the data in graphs and tables and download the reports in PDF format if required.

[1297] This frees sales representatives from manual work, enabling more efficient sales activities and achieving sustainable inventory management across the entire company.

[1298] The processing flow will be explained below.

[1299] Step 1:

[1300] The server retrieves past sales data from an internal database, which stores information such as sales date, product ID, sales quantity, price, and area.

[1301] Step 2:

[1302] The server obtains weather data from the weather agency's API, including past temperature, precipitation, wind speed, and air pressure.

[1303] Step 3:

[1304] The server retrieves economic indicator data from economic databases or trusted government APIs, such as GDP growth, unemployment rate, and consumer confidence index.

[1305] Step 4:

[1306] The server aggregates the collected sales, weather, and economic data, making the data set available in a consistent manner.

[1307] Step 5:

[1308] The server preprocesses the integrated data, imputing missing values, removing noise, and scaling the data. For example, if there is missing sales data, it is imputed using similar data from the past.

[1309] Step 6:

[1310] The server normalizes the data by scaling each variable to a range between 0 and 1. This step is important for improving the accuracy of training machine learning models.

[1311] Step 7:

[1312] The server uses the preprocessed data to train machine learning models, for example, using random forest regression or LSTM (long short-term memory) networks.

[1313] Step 8:

[1314] The server uses validation data to optimize model parameters during training and prevent overfitting or underfitting, a process that evaluates how well the model performs on real-world data.

[1315] Step 9:

[1316] The server compares the predictions with actual sales data to evaluate the accuracy of the model and tunes the model as needed, a step that is important for continuously improving the model's accuracy.

[1317] Step 10:

[1318] A user enters a request into a terminal, such as "Sales forecast for October 2023," which starts the forecast generation process.

[1319] Step 11:

[1320] The server retrieves the latest weather and economic indicator data in real time and inputs this into the trained model.

[1321] Step 12:

[1322] The server generates a sales forecast, which includes projected sales figures, peak demand times, and geographical details.

[1323] Step 13:

[1324] The server also generates inventory management recommendations based on the generated forecasts, helping to prevent overstocks and out-of-stock situations.

[1325] Step 14:

[1326] The server sends the prediction results and inventory management suggestions to the user's terminal.

[1327] Step 15:

[1328] The device visually displays the received prediction results on a dashboard, allowing users to view the results in graph and table format.

[1329] Step 16:

[1330] Users can check the forecast results on the dashboard and download the report in PDF format as needed, which allows them to develop specific sales strategies and inventory management plans.

[1331] As described above, the system collects data, preprocesses it, trains models, generates predictions, and displays the results step by step, supporting efficient sales activities and sustainable inventory management.

[1332] Example 1

[1333] 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."

[1334] Conventional sales forecasting systems often make predictions based solely on past sales data, without taking into account multiple factors such as weather data and economic indicators, resulting in low accuracy. Another issue is that few systems provide sales forecast results in real time, and the effects of missing data and noise cannot be fully mitigated. Furthermore, there is a lack of functionality that allows users to easily check sales forecast results and provides them in a format that can be used as reference material. These issues make it difficult to improve sales efficiency and protect the environment.

[1335] 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.

[1336] In this invention, the server includes: means for periodically collecting past sales data, weather data, and economic indicator data from an internal company database or an external API; means for preprocessing the collected data, filling in missing values ​​with past approximate data, removing noise, scaling, and normalizing the data; means for training a machine learning model based on the preprocessed data to learn past sales patterns and environmental factors; means for inputting the latest weather data and economic indicator data in response to a sales forecast request from a user and generating a sales forecast in real time; and means for transmitting the generated sales forecast results to the user's device, visually displaying them on a dashboard, and making them available for download as a PDF report if necessary. This improves the accuracy of sales forecasts and enables real-time information provision and visual display of results. As a result, sales efficiency and sustainable inventory management can be achieved.

[1337] An "in-house database" is a system installed within a company that stores and manages information, and holds business data such as sales data and customer information.

[1338] An "external organization's API" is an API (Application Programming Interface) provided by an external organization or service, and is a means of programmatically obtaining information such as weather data and economic indicator data.

[1339] "Sales data" refers to numerical information about sales recorded by a company during a specific period, such as the sales quantity and sales amount of a product.

[1340] "Weather data" is numerical information about the weather in a specific region or period, such as temperature, precipitation, humidity, and wind speed.

[1341] "Economic indicator data" refers to various statistical data that show the state of the economy, such as GDP growth rate, unemployment rate, and consumer price index.

[1342] "Preprocessing" refers to the preparation of data prior to data analysis or machine learning, and specifically includes processes such as filling in missing values, removing noise, scaling, and normalization.

[1343] "Missing values" refers to a state in which some data is missing, including blanks and NULL values ​​in a dataset.

[1344] "Noise" refers to unnecessary disturbances and errors contained in data that can reduce the accuracy of analyses and models.

[1345] "Scaling" is a conversion technique used to fit data values ​​into a range, usually normalizing the data range to 0 to 1 using minimum and maximum values.

[1346] "Normalization" refers to a transformation process to make the distribution of data uniform, particularly by confining the data to a specific range through scaling.

[1347] A "machine learning model" is a set of algorithms that learn patterns and rules from data and make predictions or classifications. Examples include random forests and LSTM.

[1348] "Training" is the process by which a machine learning model uses data to learn optimal parameters.

[1349] A "sales forecast request" is a request sent by a user to a server to forecast future sales based on a specific period and conditions.

[1350] "Real-time" refers to a state in which data is processed and displayed immediately at the moment it is generated or acquired.

[1351] A "dashboard" is a visual interface that allows users to see important information at a glance, displaying data in the form of graphs and tables.

[1352] "PDF format" stands for Portable Document Format, a file format for storing and sharing information in a fixed-layout format.

[1353] A "report" refers to the analysis results or report on specific data, and is often generated in PDF format or similar.

[1354] This invention is a system that collects past sales data, weather data, and economic indicator data from in-house databases and APIs of external institutions, and uses this data to perform sales forecasts using machine learning models. This system consists of a server, user terminals, and a network infrastructure that connects these.

[1355] Hardware and Software Configuration

[1356] server

[1357] The server is a computer system equipped with a high-performance processor and large memory capacity, and is installed with software for database management systems, API handlers, and machine learning models, including pandas, scikit-learn, and TensorFlow for data analysis.

[1358] Terminal

[1359] Users access the system through a web browser on devices such as PCs and tablets. The device must have an internet connection and a browser (e.g., Google Chrome or Mozilla Firefox) to view dashboards and reports.

[1360] Data collection

[1361] The server periodically connects to the company's internal database and retrieves sales data for the past five years, which is retrieved using SQL queries.

[1362] The server retrieves weather data for the past five years through the API of an external weather agency. This API request uses the HTTP GET method to retrieve data such as temperature and precipitation for a specified period.

[1363] The server uses APIs provided by governments and economic forums to collect economic indicator data (e.g., GDP growth rate, unemployment rate) for the past five years.

[1364] Data Preprocessing

[1365] The server uses the pandas library to fill in missing values ​​in the acquired sales data with similar historical data, and applies appropriate filtering techniques to remove noise.

[1366] Next, we use scikit-learn's MinMaxScaler to scale and normalize the sales and weather data to the range 0 to 1, so that the data can be fed into the machine learning model in a unified format.

[1367] Model learning

[1368] The server trains a machine learning model based on the preprocessed data. Algorithms such as random forest regression models and LSTM are used, and the model is run using scikit-learn and TensorFlow. During the training process, the dataset is split into training data and validation data, and cross-validation is applied to ensure the accuracy of the model.

[1369] Forecast Generation

[1370] When a user submits a sales forecast request from their device, they enter a prompt, such as "Sales forecast for October 2023." In response, the server again retrieves the latest weather and economic data and inputs it into the trained model.

[1371] The model generates results by predicting future sales in real time based on new data.

[1372] Results display

[1373] The server organizes the generated prediction results and sends them to the user's device, where they are visually displayed in graphs and tables on a dashboard for the user to review.

[1374] If desired, users can download the report in PDF format, a feature that uses libraries such as matplotlib and ReportLab.

[1375] Specific examples

[1376] Data collection

[1377] The server retrieves sales data for the past five years from a company database, collects weather data from the Japan Meteorological Agency's API, and obtains economic indicator data from the government's economic forum.

[1378] Data Preprocessing

[1379] The server fills in missing sales data with similar data from the past and scales the weather data to normalize it to the range 0 to 1.

[1380] Model learning

[1381] The server trains a random forest regression model to learn patterns from the data.

[1382] Forecast Generation

[1383] When a user requests "Sales forecast for October 2023," the server inputs the latest data into the model and generates a sales forecast.

[1384] Results display

[1385] The server displays the prediction results on a dashboard and allows users to download the report in PDF format.

[1386] Prompt Sentence Examples

[1387] "Show me sales forecast for October 2023."

[1388] "Get weather data for the past five years."

[1389] "Please update your economic indicator data."

[1390] This system frees sales representatives from manual work, enabling more efficient sales activities and achieving sustainable inventory management across the entire company.

[1391] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1392] Step 1:

[1393] The server connects to the company's internal database and retrieves sales data for the past five years. This is done using an SQL query. Specifically, the server executes a query such as "SELECT FROM sales WHERE date BETWEEN '2018-01-01' AND '2022-12-31'" to extract the required data. The input is the connection information and the sales data contained in the database, and the output is the sales data for the specified period.

[1394] Step 2:

[1395] The server sends an HTTP GET request to the weather agency's API to retrieve weather data for the past five years. Specifically, it sends a request such as "GET http: / / api.weather.com / data?startDate=2018-01-01&endDate=2022-12-31" to retrieve temperature and precipitation data. The input is the API endpoint information, and the output is the weather data returned by the API.

[1396] Step 3:

[1397] The server accesses the government's economic data API to obtain economic indicator data (e.g., GDP growth rate and unemployment rate) for the past five years. Specifically, it executes "GET http: / / api.economicdata.com / indicators?startDate=2018-01-01&endDate=2022-12-31" to obtain the indicator data. The input is the API endpoint information, and the output is the obtained economic indicator data.

[1398] Step 4:

[1399] The server preprocesses the collected sales data, weather data, and economic indicator data. Specifically, it creates a data frame using the pandas library and imputes missing values ​​using methods such as linear imputation. The input is the collected raw data, and the output is a data frame with missing values ​​imputed.

[1400] Step 5:

[1401] The server scales and normalizes the preprocessed data, specifically using scikit-learn's MinMaxScaler to convert values ​​to the range 0 to 1. The input is the interpolated data frame, and the output is the scaled data.

[1402] Step 6:

[1403] The server uses the scaled data to train a machine learning model, for example, a random forest regression model using scikit-learn's RandomForestRegressor. The input is the preprocessed and scaled data, and the output is a trained machine learning model.

[1404] Step 7:

[1405] The user sends a sales forecast request from the terminal. For example, they input a prompt such as "Sales forecast for October 2023." The input is the user's request, and the output is the request data to the server.

[1406] Step 8:

[1407] The server retrieves the latest weather data and economic indicator data from the API again and inputs it into the trained model, generating a sales forecast in real time. The input is the latest weather data and economic indicator data, and the output is sales forecast data.

[1408] Step 9:

[1409] The server organizes the generated sales forecast results and sends them to the user's device. The data is displayed visually on a dashboard. The input is the sales forecast data, and the output is the graphs and tables displayed on the dashboard.

[1410] Step 10:

[1411] Users can check the forecast results through a dashboard and download them in PDF format if necessary. The server generates reports using libraries such as matplotlib and ReportLab. The input is the user's download request and sales forecast data, and the output is a PDF report.

[1412] (Application example 1)

[1413] 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."

[1414] In traditional manufacturing, demand forecasting and inventory management are often performed manually, resulting in problems such as human error and reduced efficiency. It is also difficult to flexibly adjust manufacturing processes based on real-time data from the field, making it difficult to maintain a balance between supply and demand. Furthermore, it is difficult to make accurate predictions that take environmental fluctuations (weather and economic conditions) into account, leading to inventory surpluses and shortages. This calls for a method to improve production efficiency, minimize resource waste, and establish a sustainable production system.

[1415] 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.

[1416] In this invention, the server includes: means for collecting past sales data, weather data, and economic index data; means for preprocessing the collected data, imputing missing values, and normalizing the data; means for training a machine learning model based on the preprocessed data; means for inputting the latest data and generating a sales forecast in response to a sales forecast request from a user; means for displaying the generated sales forecast results on the user's terminal and visually showing them on a dashboard; and means for instructing a robot to adjust the manufacturing process based on the generated forecast results. This allows the factory's manufacturing process to be automatically adjusted in real time based on supply and demand forecasts, not only improving production efficiency but also reducing human error and enabling sustainable inventory management.

[1417] "Past sales data" refers to sales performance information recorded by companies and stores to date, specifically information such as product sales quantity, sales amount, and sales period.

[1418] "Weather data" refers to meteorological information such as temperature, precipitation, and wind speed obtained from meteorological agencies and used in forecasting models.

[1419] "Economic indicator data" refers to data that shows economic trends, and includes GDP growth rate, unemployment rate, consumer price index, etc.

[1420] "Means of collection" refers to the methods and technologies used to obtain the necessary information from various data sources.

[1421] "Preprocessing means" refers to methods and techniques for converting acquired data into an analyzable format, and specifically includes imputing missing values ​​and normalizing data.

[1422] "Methods for imputing missing values" are methods for completing incomplete data by using similar data from the past or average values, etc.

[1423] "Normalization" is a method of scaling data values ​​into a certain range so that machine learning algorithms can learn efficiently.

[1424] A "means for training a machine learning model" is a method for training a machine learning algorithm using preprocessed data to build a predictive model.

[1425] The "means for generating sales forecasts" is a method for inputting the latest data into a trained machine learning model to predict future sales.

[1426] A "user's terminal" is a device such as a PC or smartphone that the user uses to check the prediction results.

[1427] A "dashboard" refers to an interface that allows users to visually check prediction results, displaying information in the form of graphs and tables.

[1428] "Means for issuing instructions to robots to adjust the manufacturing process based on the generated prediction results" refers to methods and technologies for issuing specific operating instructions to manufacturing robots based on predicted supply and demand data, automatically adjusting the production line.

[1429] The present invention is a system that collects past sales data, weather data, and economic indicator data, and uses a machine learning model based on this data to forecast sales, and this system has been applied to factory robots. The system is configured as follows.

[1430] System Overview

[1431] This system consists of a server, user terminals, factory robots, and a network infrastructure that connects them. The server has multiple functions, including data collection, preprocessing, model learning, prediction generation, and result display. Meanwhile, users input requests through terminals or consoles and check prediction results. Furthermore, the factory robots automatically adjust the manufacturing process based on this prediction data.

[1432] Hardware and software used

[1433] Hardware:

[1434] Factory robots (e.g. Fanuc, ABB, KUKA, etc.)

[1435] Servers (cloud or on-site data center)

[1436] Sensors (monitor temperature, humidity, vibration, etc.)

[1437] software:

[1438] Machine learning frameworks (e.g. TensorFlow, PyTorch)

[1439] Database (e.g. MySQL, PostgreSQL)

[1440] API services (e.g. OpenWeatherMap, economic data API)

[1441] Robot control software (e.g. ROS - Robot Operating System)

[1442] Data processing and calculation

[1443] 1. The server retrieves past sales data from the company's internal database, collects weather data from the weather agency's API, and collects economic indicator data from the economic database.

[1444] 2. The server preprocesses the collected data. Specifically, it imputes missing values, removes noise, and scales and normalizes the data. For example, if sales data is missing for a period, it is imputed using similar data from the past.

[1445] 3. The server uses the preprocessed data to train a machine learning model. This model uses a selected algorithm (e.g., random forest regression or LSTM) for sales forecasting. The model learns sales patterns based on past data.

[1446] 4. When a user sends a sales forecast request from their device, the server retrieves the latest weather data and economic trend data from the API and uses this data to generate a sales forecast in real time.

[1447] 5. The server generates the prediction results and sends them to the user's device. The user can view the results visually on a dashboard and optionally download a report in PDF format.

[1448] 6. Based on the generated prediction results, the server sends instructions to the factory robot to adjust the manufacturing process, allowing the factory robot to automatically optimize the manufacturing process based on the supply and demand balance.

[1449] Specific examples

[1450] Data collection:

[1451] The server retrieves the production and inventory data for the past two years from the factory database, and also uses APIs to retrieve weather data (e.g., temperature, humidity) for the past two years, and collects economic data (e.g., GDP growth rate, production index).

[1452] Data preprocessing:

[1453] The server fills in missing data using past approximate data, scales sales data and weather data to a range of 0 to 1, and converts them into a format suitable for the model.

[1454] Model training:

[1455] The server trains a random forest regression model on the preprocessed data, which is used to learn about past sales patterns and the effects of weather and economic conditions and to predict future sales.

[1456] Forecast generation:

[1457] When a user requests "Sales forecast for October 2023," the server retrieves the latest weather forecast and economic indicators and inputs them into the trained model, which then outputs detailed sales forecasts and supply and demand plans for each region.

[1458] Results display and automatic adjustment:

[1459] The server then organizes the generated sales forecasts and displays them on the user's dashboard. Based on the forecasts, the server sends instructions to factory robots to adjust the manufacturing process, for example, adjusting the production speed of a specific product line to prevent overstocks or shortages.

[1460] Example prompts to input to the generative AI model

[1461] Input historical production data, weather data, and economic indicator data to generate a production demand forecast for the next month.

[1462] Using this system, supply and demand forecasts and production process optimization at manufacturing sites can be performed in real time, improving production efficiency and optimizing inventory management.

[1463] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1464] Step 1: Data collection

[1465] The server retrieves past sales data from the company's internal database, and uses APIs to collect weather data from meteorological agencies and economic indicator data from economic databases. As a result, past sales data, weather data, and economic indicator data are aggregated on the server as input data. Specifically, this includes sales quantity, sales amount, temperature, precipitation, GDP growth rate, unemployment rate, etc.

[1466] Step 2: Data Preprocessing

[1467] The server then fills in missing values ​​in the collected data, removes noise, and scales and normalizes the data. Specifically, it fills in missing sales data using similar historical data, and does the same for weather data and economic indicator data. The result is a preprocessed dataset that can be used by machine learning models.

[1468] Step 3: Train the machine learning model

[1469] The server uses the preprocessed data to train a machine learning model for sales forecasting. Specifically, it uses algorithms such as random forest regression and LSTM to learn sales patterns and correlations with various metrics, resulting in a trained machine learning model.

[1470] Step 4: Receive a prediction request

[1471] The user sends a sales forecast request from the terminal. For example, they enter a prompt sentence such as "Please generate sales forecast for October 2023." This becomes the input data to the server.

[1472] Step 5: Generate forecasts

[1473] The server receives the request from the user and retrieves the latest weather and economic indicator data from the API again. The retrieved data is input into the trained model to generate a sales forecast. This operation outputs real-time sales forecast data.

[1474] Step 6: View the results

[1475] The server sends the generated sales forecast results to the user's device, where the user can visually check the forecast results on the dashboard of their device. For example, the sales forecast data can be displayed in graph or table format.

[1476] Step 7: Adjust the manufacturing process based on the results

[1477] Based on the generated predictions, the server sends instructions to the factory robots to adjust the manufacturing process. The factory robots then use the instructions received from the server to optimize the manufacturing process in real time, for example, by adjusting the production speed of a particular product line or changing the amount of materials required.

[1478] Through these processing steps, the system automates everything from sales forecasts to real-time adjustments to manufacturing processes, enabling efficient production management.

[1479] 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.

[1480] This invention combines a system that collects past sales data, weather data, and economic indicator data and uses machine learning models to forecast sales, with an emotion engine that recognizes user emotions. This system is designed to improve sales efficiency and protect the environment, and also enables flexible responses based on user emotions.

[1481] System Overview

[1482] This system consists of a server, a user's device, an emotion engine, and a network infrastructure that connects these components. The server has multiple functions, including data collection, preprocessing, model training, prediction generation, result display, and emotion recognition. Meanwhile, users input requests through their device and check prediction results.

[1483] Program processing flow

[1484] 1. Data Collection

[1485] The server periodically collects historical sales data, weather data, and economic indicator data. The sales data is obtained from an internal database, the weather data is obtained from a weather agency's API, and the economic indicator data is obtained from an economic database.

[1486] 2. Data Preprocessing

[1487] The server preprocesses the collected data by filling in missing values, removing noise, scaling, and normalizing it. For example, if there are gaps in sales data, it fills them in using similar data from the past.

[1488] 3. Model Training

[1489] The server uses the preprocessed data to train a machine learning model using techniques such as random forest regression and LSTM (long short-term memory) networks. The model learns sales patterns based on past data.

[1490] 4. Emotion recognition

[1491] The device analyzes the user's facial expressions and tone of voice, and uses an emotion engine to recognize the user's emotions. The emotion data is sent to the server in real time.

[1492] 5. Prediction Generation

[1493] When a user sends a sales forecast request from their device, the server retrieves the latest weather and economic indicator data, inputs this data into the trained model, and adjusts the forecast taking into account emotion data from the emotion engine.

[1494] 6. Results display

[1495] The server generates prediction results and inventory management suggestions and sends them to the user's device. The user can view the results visually on a dashboard, and emotion data is used to dynamically adjust the display content and interface.

[1496] Specific examples

[1497] Data collection

[1498] The server retrieves sales data for the past five years from an internal database, and also retrieves weather data (temperature, precipitation, etc.) for the past five years from meteorological agencies via API, as well as economic indicator data (GDP growth rate, unemployment rate).

[1499] Data Preprocessing

[1500] The server normalizes the collected data by imputing missing values, removing noise, and scaling. Missing sales data is interpolated based on similar patterns in the past.

[1501] Model learning

[1502] The server trains a random forest regression model on the pre-processed data to learn about past sales patterns and the effects of weather and economic conditions.

[1503] emotion recognition

[1504] The device analyzes the user's facial expressions and voice tone in real time and obtains emotional data through an emotion engine, which can determine whether the user is, for example, stressed or relaxed.

[1505] Forecast Generation

[1506] When a user requests "Sales forecast for October 2023," the server retrieves the latest weather forecast, economic indicator data, and sentiment data, and inputs these into the model to generate a sales forecast. If the server determines that the user is stressed, it will carefully provide feedback on the forecast and present the results.

[1507] Results display

[1508] The server sends the forecast results and inventory management suggestions to the user's device, which displays them in graph and table format on a dashboard. The user can check the results and download the report in PDF format if necessary. The display content is also adjusted and the interface is customized based on the user's sentiment.

[1509] In this way, the present invention can provide more accurate sales forecasts and inventory management while responding to user emotions.

[1510] The processing flow will be explained below.

[1511] Step 1:

[1512] The server retrieves past sales data from an internal database, which stores information such as sales date, product ID, sales quantity, price, and area.

[1513] Step 2:

[1514] The server obtains weather data from the weather agency's API, specifically, past temperature, precipitation, wind speed, air pressure, etc.

[1515] Step 3:

[1516] The server retrieves economic indicator data from economic databases or trusted government APIs, such as GDP growth, unemployment rate, and consumer confidence index.

[1517] Step 4:

[1518] The server aggregates the collected sales data, weather data, and economic indicator data, allowing for a consistent analysis of the data set.

[1519] Step 5:

[1520] The server preprocesses the integrated data by imputing missing values, removing noise, and scaling it. For example, if there is missing sales data, it fills in the missing data using similar data from the past.

[1521] Step 6:

[1522] The server uses the preprocessed data as training data for machine learning models, such as random forest regression and LSTM (long short-term memory) networks, to train the models.

[1523] Step 7:

[1524] The server uses validation data to optimize the parameters of the trained model and prevent overfitting and underfitting.

[1525] Step 8:

[1526] The device analyzes the user's facial expressions and voice tone in real time and recognizes their emotions using an emotion engine. The recognized emotion data is then sent to the server.

[1527] Step 9:

[1528] When a user requests "Sales forecast for October 2023" from their device, the server retrieves the latest weather data, economic indicator data, and sentiment data.

[1529] Step 10:

[1530] The server then inputs this data into a trained model to generate sales forecasts, with the sentiment data being used to fine-tune the results.

[1531] Step 11:

[1532] The server organizes the generated sales forecast results and simultaneously creates inventory management proposals, thereby preventing excess inventory and out-of-stock situations.

[1533] Step 12:

[1534] The server transmits the sales forecast results and inventory management proposals to the user's terminal.

[1535] Step 13:

[1536] The terminal displays the received sales forecast results and inventory management suggestions on a dashboard, including graphs and tables.

[1537] Step 14:

[1538] Users can check the forecast results on the dashboard and download the report in PDF format if necessary.

[1539] Step 15:

[1540] The device dynamically adjusts the display content and interface based on the user's emotions. For example, if the user is feeling stressed, the device will change the display to a simpler display and a more visually soothing color scheme.

[1541] As described above, the system performs data collection, preprocessing, model training, emotion recognition, prediction generation, and result display at each step, supporting efficient sales activities and sustainable inventory management while also improving the user experience.

[1542] Example 2

[1543] 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."

[1544] Conventional sales forecasting systems typically collect past sales data, weather data, economic indicator data, etc., and use machine learning to forecast sales. However, they were unable to take into account individual factors such as the user's emotional state. This meant that they were unable to respond flexibly to user emotions, making it difficult to improve forecast accuracy and user experience. Furthermore, there were limited ways to visually confirm forecast results and present them in a format that was easy for users to understand.

[1545] 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.

[1546] In this invention, the server includes: means for collecting past sales data, weather data, and economic indicator data; means for preprocessing the collected data, imputing missing values, and normalizing the data; means for training a machine learning model based on the preprocessed data; means for recognizing user emotion data and transmitting it to the server; means for inputting the latest data and emotion data and generating a sales forecast in response to a sales forecast request from the user; and means for displaying the generated sales forecast results on the user's terminal and visually presenting them on a dashboard. This enables sales forecasts that take the user's emotional state into account, improving forecast accuracy and the user experience. Furthermore, visually displaying the forecast results makes it possible to provide a system that is easy for users to understand and use.

[1547] "Past sales data" refers to records of past sales and purchases of goods conducted by a company, and includes data such as sales quantities, amounts, and dates.

[1548] "Weather data" refers to data including meteorological information such as temperature, precipitation, wind speed, and humidity obtained from meteorological agencies.

[1549] "Economic indicator data" is statistical data used to understand the state of the economy, such as gross domestic product (GDP), unemployment rate, and inflation rate.

[1550] "Preprocessing" refers to the processing performed on collected raw data, and involves data cleansing and normalization, such as filling in missing values, removing noise, and scaling.

[1551] A "machine learning model" is an algorithm or computational model that learns patterns and regularities from given data and makes predictions and judgments about future data.

[1552] "Emotion data" is data that indicates the emotional state of the user analyzed from facial expressions, tone of voice, and the like.

[1553] A "server" is a central computer that performs data collection, preprocessing, model training, prediction generation, and display of results.

[1554] A "terminal" is a computer or mobile device used by a user to enter data and confirm results.

[1555] A "dashboard" is an interface that allows users to visually check information such as sales forecast results and inventory management proposals.

[1556] This invention is a system that collects and analyzes past sales data, weather data, economic indicator data, and even user emotion data, and uses machine learning models to forecast sales. This system is composed of a server, a terminal, an emotion recognition engine, and a network infrastructure that links these components. Below, we will explain in detail how each component works.

[1557] Data collection

[1558] The server periodically collects past sales data, weather data, and economic indicator data. Sales data is obtained from an internal database, weather data is obtained using the weather agency's API, and economic indicator data is obtained from an economic database. Specifically, sales data is extracted from the database by executing SQL queries using Python. Weather data is obtained in JSON format from the weather agency's API using an HTTP request, for example, and economic indicator data is collected from the IMF and OECD APIs.

[1559] Data Preprocessing

[1560] The server preprocesses the collected data. This preprocessing includes imputing missing values, removing noise, scaling, and normalizing the data. For example, it uses Python's pandas library to impute missing values ​​in the data frame with the mean value and scales the data using scikit-learn's StandardScaler. This cleans the data so that the machine learning model can operate optimally.

[1561] Model learning

[1562] The server uses the preprocessed data to train a machine learning model. The main techniques used are random forest regression and LSTM (long short-term memory). For example, a model is created using scikit-learn's RandomForestRegressor and trained using the training data. The server also evaluates the accuracy of the model and adjusts hyperparameters using cross-validation if necessary.

[1563] emotion recognition

[1564] The device captures the user's facial expressions and voice tone and analyzes them with an emotion recognition engine. For example, it uses OpenCV to analyze video from a webcam and recognize emotions from facial expressions. It also uses the Librosa library to extract voice features and classify emotions using a deep learning model. This emotion data is sent to a server in real time and incorporated into a prediction algorithm.

[1565] Forecast Generation

[1566] When a user requests a sales forecast through their device, the server retrieves the latest weather data, economic indicator data, and emotion data and inputs this into the trained model. For example, if a user requests "Sales forecast for October 2023," the server collects the data and generates a sales forecast. The user's emotion data is also taken into account and the forecast results are adjusted. For example, a more conservative forecast is provided to a user who is feeling stressed.

[1567] Results display

[1568] The server generates forecast results and inventory management proposals and sends them to the user's device. The device visually displays the results on a dashboard, allowing the user to intuitively understand the results. For example, graphs and tables are created using the Plotly library and provided to the user. The user can check the results and download the report in PDF format if necessary. The color and layout of the user interface are also dynamically adjusted based on the user's mood.

[1569] Example prompts to input to the generative AI model

[1570] "Please provide a sales forecast for October 2023. Please use the latest weather and economic indicator data. Please also take into account the user's emotional state and provide results."

[1571] In this way, the present invention enables sales forecasting that takes into account the emotional state of the user, thereby realizing a system that provides higher accuracy and a better user experience.

[1572] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1573] Step 1: Data collection

[1574] The server collects historical sales data, weather data, and economic indicator data.

[1575] Input: In-house databases, weather agency APIs, economic databases.

[1576] What it does: Uses Python to run SQL queries to extract sales data. Uses HTTP requests to retrieve weather data from meteorological agency APIs and economic indicators data from IMF and OECD APIs. Saves this data in CSV and JSON formats.

[1577] Output: Files with historical sales data, weather data, and economic indicator data.

[1578] Step 2: Data Preprocessing

[1579] The server pre-processes the collected data.

[1580] Inputs: Historical sales data, weather data, economic indicator data.

[1581] Specific operation: Uses the pandas library to impute missing values ​​in a data frame with the mean and remove outliers. Scales the data using scikit-learn's StandardScaler, which imputes missing values, removes noise, and scales the data.

[1582] Output: Preprocessed data.

[1583] Step 3: Model training

[1584] The server trains a machine learning model using the preprocessed data.

[1585] Input: Preprocessed data.

[1586] Specific operation: Create a model using scikit-learn's RandomForestRegressor, train the model using the training data, evaluate the model's accuracy using cross-validation, and adjust the hyperparameters as needed.

[1587] Output: A trained machine learning model.

[1588] Step 4: Emotion Recognition

[1589] The terminal recognizes the user's emotion data.

[1590] Input: User's facial expressions and vocal tone.

[1591] Specific operation: Analyzes video from a webcam using OpenCV and recognizes emotions from facial expressions. Extracts audio features using the Librosa library and classifies emotions using a deep learning model. Sends emotion data to the server in real time using WebSocket.

[1592] Output: Real-time sentiment data.

[1593] Step 5: Generate forecasts

[1594] The user submits a sales forecast request through a terminal.

[1595] Inputs: Sales forecast request from user, weather updates, economic indicators, and sentiment data.

[1596] How it works: The server receives the request and inputs the latest weather data, economic indicators, and sentiment data into the model. It then adjusts the prediction results to account for the user's emotional state and generates a sales forecast.

[1597] Output: Sales forecast results.

[1598] Step 6: View the results

[1599] The server generates the forecast results and inventory management proposals and sends them to the user's terminal.

[1600] Input: Sales forecast results, inventory management proposals.

[1601] Specific operation: The server generates forecast results and inventory management suggestions in JSON format and sends them to the terminal. The terminal uses the Plotly library to visually display them as graphs and tables on a dashboard. The UI color and layout are dynamically adjusted according to the user's mood. The user can also download the results in PDF format.

[1602] Output: Forecast results displayed in a dashboard, a report in PDF format.

[1603] (Application example 2)

[1604] 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."

[1605] Conventional sales forecasting systems make predictions based on past data and environmental data, but because they do not take user emotions into account, they often lack the accuracy of predictions and flexibility in customer service. Furthermore, in customer service situations in physical stores, there is no way to accurately grasp customer emotions, making it difficult to efficiently manage inventory and recommend appropriate products.

[1606] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting past sales data, weather data, and economic index data, means for preprocessing the collected data, imputing missing values, and normalizing the data, means for training a machine learning model based on the preprocessed data, means for recognizing user emotions in real time and using the emotion data, means for inputting the latest data and emotion data and generating a sales forecast in response to a sales forecast request from a user, and means for displaying the generated sales forecast results on the user's terminal and visually showing them on a dashboard. This enables flexible sales forecasting and inventory management based on user emotions, as well as customer service in physical stores.

[1607] "Historical sales data" refers to sales information recorded based on previous sales activity.

[1608] "Weather data" refers to numerical data or information that represents meteorological information, and is data that indicates the weather conditions at a specific date, time, and location.

[1609] "Economic indicator data" is data based on multiple economic indicators that show economic trends and conditions.

[1610] "Preprocessing" is the process of preparing collected data using methods such as filling in missing values, removing noise, and scaling.

[1611] A "machine learning model" is an algorithm that learns patterns and rules from large amounts of data and performs tasks such as prediction and classification.

[1612] "User emotion data" is information that indicates the user's emotional state in real time, obtained from the user's facial expressions, tone of voice, and the like.

[1613] An "emotion engine" is a technology or device that analyzes input data such as a user's facial expressions and voice to determine their emotional state.

[1614] A "sales forecast request" is a request made by a user to the system to forecast future sales.

[1615] "Latest data" refers to the most recent information or data available at the present time.

[1616] "Sales forecasting" refers to the use of machine learning models and other data to estimate future sales.

[1617] A "dashboard" is an interface designed for data visualization and analysis, and displays multiple pieces of information in an integrated manner.

[1618] "Inventory management proposal" is a method of proposing appropriate inventory levels and product placement based on past data and the latest information.

[1619] "PDF" is an abbreviation for Portable Document Format, a standard format for electronically storing and sharing documents.

[1620] This invention combines a system that collects past sales data, weather data, and economic indicator data and uses machine learning models to forecast sales, with an emotion engine that recognizes user emotions. This system is designed to improve sales efficiency and protect the environment, and also enables flexible responses based on user emotions.

[1621] The server comprises the following means:

[1622] 1. A means of collecting historical sales data, weather data, and economic indicator data

[1623] 2. Preprocessing the collected data, imputing missing values, and normalizing them

[1624] 3. A means to train machine learning models on preprocessed data

[1625] 4. Real-time recognition of user emotions and the use of emotional data

[1626] 5. A means of generating a sales forecast by inputting the latest data and sentiment data in response to a sales forecast request from a user.

[1627] 6. A means to display the generated sales forecast results on the user's device and visually display them on a dashboard

[1628] This will enable flexible sales forecasting and inventory management based on user sentiment, as well as customer service in physical stores.

[1629] Hardware and software used

[1630] Hardware

[1631] 1. Server: Performs key processing such as data collection, preprocessing, model training, and prediction generation.

[1632] 2. Smart glasses: Devices that detect the user's facial expressions and vocal tone to analyze emotions.

[1633] software

[1634] 1. OpenCV: An open-source library used for face detection and image processing.

[1635] 2. TensorFlow: A platform used to analyze facial expressions and recognize emotions using deep learning models.

[1636] 3. Scikit-Learn: A library used to perform sales forecasting using random forest regression models.

[1637] Example of a system

[1638] 1. Data Collection:

[1639] The server retrieves sales data for the past five years from an internal database, and also retrieves weather data (temperature, precipitation, etc.) for the past five years from meteorological agencies via API, as well as economic indicator data (GDP growth rate, unemployment rate).

[1640] 2. Data Preprocessing:

[1641] The server normalizes the collected data by imputing missing values, removing noise, and scaling. Missing sales data is imputed based on similar patterns in the past.

[1642] 3. Model training:

[1643] Using the pre-processed data, the server trains a random forest regression model to learn about past sales patterns and the effects of weather and economic conditions.

[1644] 4. Emotion recognition:

[1645] The smart glasses analyze the user's facial expressions and vocal tone in real time and use TensorFlow to obtain emotional data, which can determine whether the user is stressed or relaxed, for example.

[1646] 5. Sales forecast generation:

[1647] When a user requests "Sales forecast for October 2023," the server retrieves the latest weather forecast, economic indicator data, and sentiment data, and inputs these into the model to generate a sales forecast. If the server determines that the user is stressed, it will carefully provide feedback on the forecast and present the results.

[1648] 6. Results display:

[1649] The server sends the forecast results and inventory management suggestions to the user's device, which then visually displays them on a dashboard. The display content is also adjusted and the interface is customized based on the user's emotions.

[1650] Prompt Sentence Examples

[1651] "Local time in Tokyo on October 15, 2023, the weather is rainy, the temperature is 15 degrees, the GDP growth rate is 2.5%, the unemployment rate is 3%, and customer sentiment is stressed. Based on this data, generate a sales forecast."

[1652] This allows the present invention to leverage user sentiment and external data to optimize sales and customer experience in physical stores.

[1653] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1654] Step 1: Data collection

[1655] The server periodically collects historical sales data, weather data, and economic indicator data. The sales data is obtained from an internal company database, the weather data is obtained from a weather agency's API, and the economic indicator data is obtained from an economic database. The input is data from each data source, and the output is the collected data set.

[1656] Step 2: Data Preprocessing

[1657] The server preprocesses the collected data. Specifically, it imputes missing values ​​in the data, removes noise, and scales and normalizes it. For example, if there are gaps in sales data, it imputes them using similar data from the past. The input is the data collected in step 1, and the output is the preprocessed data.

[1658] Step 3: Model training

[1659] The server uses the preprocessed data to train a machine learning model. Training techniques such as random forest regression and LSTM (long short-term memory) networks are used. The model learns sales patterns based on past data. The input is the preprocessed data from step 2, and the output is the trained machine learning model.

[1660] Step 4: Emotion Recognition

[1661] The device analyzes the user's facial expressions and voice tone, and uses an emotion engine to recognize the user's emotions. The input is the user's voice and facial expression data acquired from the device, and the output is the analyzed emotion data.

[1662] Step 5: Generate a sales forecast

[1663] When a user sends a sales forecast request from their device, the server retrieves the latest weather and economic indicator data and inputs this data and sentiment data into the trained model. The server then adjusts the forecast and generates a sales forecast. The input is the user's request, the latest data, and sentiment data, and the output is the sales forecast result.

[1664] Step 6: View the results

[1665] The server generates forecast results and inventory management proposals and sends them to the user's terminal, where they are visually displayed as a dashboard. The input is the sales forecast results, and the output is the visually displayed data.

[1666] This process flow allows the present invention to respond to user emotions while providing more accurate sales forecasts and inventory management.

[1667] 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.

[1668] 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.

[1669] 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.

[1670] 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.

[1671] 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.

[1672] 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.

[1673] 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).

[1674] 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.

[1675] 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."

[1676] 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.

[1677] 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).

[1678] 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.

[1679] 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.

[1680] 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.

[1681] 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.

[1682] 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.

[1683] 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.

[1684] 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.

[1685] 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.

[1686] 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.

[1687] 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.

[1688] The following is further disclosed regarding the above embodiment.

[1689] (Claim 1)

[1690] A means of collecting historical sales data, weather data, and economic indicator data;

[1691] A means to preprocess the collected data, impute missing values, and normalize them;

[1692] a means for training a machine learning model on the preprocessed data; and

[1693] a means for inputting the latest data and generating a sales forecast in response to a sales forecast request from a user;

[1694] A means for displaying the generated sales forecast results on the user's device and visually showing them on a dashboard;

[1695] A system including:

[1696] (Claim 2)

[1697] 2. The system according to claim 1, further comprising means for simultaneously making inventory management suggestions based on past sales data, the latest weather data, and economic trend data.

[1698] (Claim 3)

[1699] 10. The system of claim 1, further comprising means for making the sales forecast results downloadable as a report in PDF format.

[1700] "Example 1"

[1701] (Claim 1)

[1702] A means to periodically collect past sales data, weather data, and economic indicator data from internal company databases and external APIs.

[1703] A means of preprocessing the collected data, imputing missing values ​​with historical approximate data, removing noise, scaling and normalizing it;

[1704] A means to train machine learning models based on pre-processed data to learn historical sales patterns and environmental factors;

[1705] A means for inputting the latest weather data and economic indicator data and generating a sales forecast in real time in response to a sales forecast request from a user;

[1706] A means to send the generated sales forecast results to the user's device, visually display them on a dashboard, and make them downloadable as a report in PDF format if necessary;

[1707] A system including:

[1708] (Claim 2)

[1709] 2. The system according to claim 1, further comprising means for simultaneously making inventory management suggestions based on past sales data, the latest weather data, and economic trend data.

[1710] (Claim 3)

[1711] 10. The system of claim 1, further comprising means for a user to submit a sales forecast for a particular time period or region as an input prompt based on the generative AI model.

[1712] "Application Example 1"

[1713] (Claim 1)

[1714] A means of collecting historical sales data, weather data, and economic indicator data;

[1715] A means to preprocess the collected data, impute missing values, and normalize them;

[1716] a means for training a machine learning model on the preprocessed data; and

[1717] a means for inputting the latest data and generating a sales forecast in response to a sales forecast request from a user;

[1718] A means for displaying the generated sales forecast results on the user's device and visually showing them on a dashboard;

[1719] a means for instructing a robot to adjust the manufacturing process based on the generated prediction result;

[1720] A system including:

[1721] (Claim 2)

[1722] 2. The system according to claim 1, further comprising means for simultaneously making inventory management suggestions based on past sales data, the latest weather data, and economic trend data.

[1723] (Claim 3)

[1724] 10. The system of claim 1, further comprising means for making the sales forecast results downloadable as a report in PDF format.

[1725] "Example 2: Combining Emotion Engines"

[1726] (Claim 1)

[1727] A means of collecting historical sales data, weather data, and economic indicator data;

[1728] A means to preprocess the collected data, impute missing values, and normalize them;

[1729] a means for training a machine learning model on the preprocessed data; and

[1730] means for recognizing user emotion data and transmitting it to a server;

[1731] a means for inputting the latest data and sentiment data and generating a sales forecast in response to a sales forecast request from a user;

[1732] A means for displaying the generated sales forecast results on the user's device and visually showing them on a dashboard;

[1733] A system including:

[1734] (Claim 2)

[1735] 10. The system of claim 1, further comprising means for simultaneously providing inventory management suggestions based on past sales data, latest weather data, economic indicator data, and sentiment data.

[1736] (Claim 3)

[1737] 10. The system of claim 1, further comprising means for making the sales forecast results downloadable as a report in PDF format.

[1738] "Application example 2 when combining emotion engines"

[1739] (Claim 1)

[1740] A means of collecting historical sales data, weather data, and economic indicator data;

[1741] A means to preprocess the collected data, impute missing values, and normalize them;

[1742] a means for training a machine learning model on the preprocessed data; and

[1743] a means for inputting the latest data and generating a sales forecast in response to a sales forecast request from a user;

[1744] a means for acquiring user emotion data in real time using an emotion engine and adjusting the sales forecast;

[1745] A means for displaying the generated sales forecast results on the user's device and visually showing them on a dashboard;

[1746] A system including:

[1747] (Claim 2)

[1748] 2. The system according to claim 1, further comprising means for simultaneously making inventory management suggestions based on past sales data, the latest weather data, economic trend data, and user emotion data.

[1749] (Claim 3)

[1750] 10. The system of claim 1, further comprising means for making the sales forecast results downloadable as a report in PDF format. [Explanation of symbols]

[1751] 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 sales data, weather data, and economic indicator data; A means to preprocess the collected data, impute missing values, and normalize them; a means for training a machine learning model on the preprocessed data; and a means for inputting the latest data and generating a sales forecast in response to a sales forecast request from a user; A means for displaying the generated sales forecast results on the user's device and visually showing them on a dashboard; A system including:

2. 2. The system according to claim 1, further comprising means for simultaneously making inventory management suggestions based on past sales data, the latest weather data, and economic trend data.

3. The system according to claim 1 , further comprising means for making the sales forecast results downloadable as a report in PDF format.

Images