System

Abstract

A system is provided.SOLUTION: A system comprising: a terminal collecting data from a communication tool and a material used by a user; the terminal uploading the collected data to a server; the server preprocessing the received data; the server evaluating a risk score associated with an insider trade using a machine learning model based on the preprocessed data; the server aggregating a plurality of risk scores to determine a final risk level; the server recommending a safe trade brand to the user based on the risk level; and the terminal notifying the user of a recommendation result.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] Insider trading poses a high legal risk in corporate activities, and there is a possibility that this risk may be inflicted unintentionally. In particular, when people within a company come across information related to stock trading in the course of their daily communications, it is extremely difficult to determine whether the content constitutes insider information. As a result, employees themselves must always be aware of the risks of insider trading when making trades, which places a significant psychological burden. There is a need for a system that can alleviate this burden while avoiding legal risks. [Means for solving the problem]

[0005] The present invention provides a system including: a terminal for collecting data from communication tools and materials used by a user; a terminal for uploading the collected data to a server; a server for preprocessing the received data; a server for evaluating a risk score related to insider trading using a machine learning model based on the preprocessed data; a server for aggregating multiple risk scores and determining a final risk level; a server for recommending safe stocks to the user based on the risk level; and a terminal for notifying the user of the recommendation results. This system automatically evaluates the insider trading risks that employees encounter on a daily basis and enables them to identify safe stocks to buy and sell. This system can reduce the legal risks associated with insider trading and also ease the psychological burden on employees.

[0006] A "terminal" refers to a device used by a user, such as a computer or smartphone, that has the function of collecting data from communication tools and communicating with a server.

[0007] "User" refers to an individual who belongs to a company and trades stocks, and who uses this system to trade while avoiding the risk of insider trading.

[0008] "Communication tools" refers to information exchange methods that users use on a daily basis, such as email, chat apps, and conference tools.

[0009] "Materials" includes documents and digital records created or used by users in the course of their business, such as meeting minutes, calendar entries, and customer registration system trails.

[0010] "Data" refers collectively to information such as text, audio, and images obtained from user communications and materials.

[0011] "Server" refers to a centralized computer system that receives, analyzes, and processes data uploaded from terminals.

[0012] "Preprocessing" refers to processes such as standardization, tokenization, and removal of stop words that are carried out to format collected data so that it is easier to analyze.

[0013] A "machine learning model" is an algorithm trained on large amounts of data and used to predict risk scores for new data.

[0014] A "risk score" is a number that indicates the degree of risk associated with insider trading in data, and is calculated using a machine learning model.

[0015] The "risk level" indicates an overall assessment of insider trading risk obtained by aggregating multiple risk scores.

[0016] "Trading stock" refers to the stock that a user is trading, and is the target for the system to evaluate insider trading risk.

[0017] "Recommendation results" refers to a list of stocks that can be traded that the server presents to the user based on risk level. [Brief explanation of the drawings]

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

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

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

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

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

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

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

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

[0026] [First embodiment]

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

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

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

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

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

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

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

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

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

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

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

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

[0039] The system of the present invention, in which terminals and a server work in cooperation with each other, supports users in safely trading stocks. This system operates in the following manner: the user's terminal collects data from communication tools and documents, uploads it to the server, the server preprocesses the data, performs risk assessment, and finally recommends safe stocks to the user.

[0040] Data collection and upload

[0041] The device periodically collects data such as the user's emails, chat messages, calendar events, meeting minutes, etc. This collected data is temporarily stored on the device and then uploaded to the server at a specified time or trigger.

[0042] Data Preprocessing

[0043] When the server receives data sent from the terminal, it first preprocesses the data. Preprocessing includes standardizing the text data, tokenizing it using natural language processing, removing stop words, etc. For example, HTML tags are removed from the email body and the email is formatted as pure text.

[0044] Risk Assessment

[0045] Based on the pre-processed text data, the server performs risk assessment using a pre-trained machine learning model, which is trained on historical insider trading data and calculates an insider trading risk score for new data.

[0046] Risk score aggregation

[0047] The risk scores calculated for each data point are aggregated by the server to determine a final risk level, which is then associated with each individual stock and provides an assessment of the stock the user plans to trade.

[0048] Stock recommendations and notifications

[0049] The server recommends safe stocks to the user based on the aggregated risk level. The results are sent to the terminal and notified to the user. For example, a list of specific stocks is displayed in the form of "These are the stocks that you can safely buy and sell today."

[0050] Specific examples

[0051] For example, if a user sends or receives an email stating, "At the next board meeting, we will discuss a new product from Company X," this email is collected by the device and uploaded to the server. The server preprocesses this email and inputs it into a machine learning model to determine that it is likely to pose a high insider trading risk. In this case, the risk score will be high, and the related Company X stock will be excluded as a non-recommended stock. Meanwhile, other low-risk stocks will be recommended to the user.

[0052] This system allows users to trade stocks safely without being aware of the risk of insider trading, and also reduces legal risks for the company as a whole. Each component of this system is built by combining existing technologies, and is widely available as a feasible means.

[0053] The processing flow will be explained below.

[0054] Step 1:

[0055] The device collects data such as the user's email, calendar, meeting minutes, and chat messages. This collection process is automated and runs periodically. Specifically, the device accesses the user's mailbox at specified intervals to retrieve new emails. Calendar and meeting minutes data is also collected in the same way.

[0056] Step 2:

[0057] The device uploads the collected data to the server. The collected data is temporarily stored in local storage and then transferred to the server according to the upload schedule. At this time, the data is encrypted for security reasons.

[0058] Step 3:

[0059] The server pre-processes the data it receives. This includes cleaning the text data. Specifically, HTML tags are removed from emails and chat messages, converting them to pure text, and then tokenizing them using natural language processing to remove meaningless words (stop words).

[0060] Step 4:

[0061] The server inputs the pre-processed text data into a machine learning model, a pre-trained risk assessment model, which outputs a score indicating the degree of risk associated with insider trading for each piece of data.

[0062] Step 5:

[0063] The server aggregates multiple risk scores to determine the final risk level. Specifically, it aggregates the risk scores obtained for each data point using statistical methods to calculate an overall risk level for each stock. This allows it to evaluate whether a particular stock can be traded.

[0064] Step 6:

[0065] The server recommends safe trading stocks to users based on their risk level, which includes a process of selecting and listing stocks with low risk scores. The selected stocks are then recommended to users.

[0066] Step 7:

[0067] The terminal notifies the user of the recommendation results sent from the server. Specifically, the terminal displays a "list of stocks that are safe to buy and sell today" via a pop-up notification or email, providing the user with information to trade safely. This notification allows the user to proceed with trading with peace of mind.

[0068] Example 1

[0069] 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."

[0070] There is a need to provide an effective support system to prevent users from unintentionally causing legal problems in stock trading, which involves risks such as market manipulation and insider trading. This system must appropriately collect data from the communication methods and information used by users, perform risk assessments based on that data, and recommend safe trading assets.

[0071] 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.

[0072] In this invention, the server includes a means for preprocessing the received data, a means for evaluating a risk score associated with market manipulation using a machine learning model based on the preprocessed data, and a means for aggregating multiple risk scores to determine a final risk level, thereby enabling the recommendation of safe trading assets for users to trade stocks with peace of mind.

[0073] A "terminal" is an electronic device used by a user, which has a means of communication and the ability to collect information and upload it to a server.

[0074] "Data collection" is the process by which the terminal obtains the necessary data from the user's communication means and information.

[0075] A "server" is a central computer system that processes data received from terminals and provides functions such as risk assessment and asset recommendations.

[0076] "Data preprocessing" is the process by which the server converts the raw data it receives into an analyzable format, and specifically includes text standardization, tokenization, and removal of commonly used words.

[0077] A "machine learning model" refers to an algorithm or mathematical model that learns patterns from data and makes predictions or classifications for new data.

[0078] A "risk score" is a numerical assessment of the risk associated with market manipulation and insider trading, assessed using a machine learning model.

[0079] The "risk level" is a comprehensive index that aggregates multiple risk scores and indicates the overall degree of risk.

[0080] "Safe trading assets" refer to assets that are deemed to have a low risk level and that users can trade with confidence.

[0081] "Recommendation" refers to the act of the server recommending a transaction to the user based on the results of the risk assessment.

[0082] "Notification" is the process by which the terminal notifies the user of the recommendation results received from the server.

[0083] MODE FOR CARRYING OUT THE INVENTION

[0084] The system of the present invention aims to help users safely trade stocks, and involves the cooperation of a terminal and a server. Specifically, the terminal collects data from the user's various communication methods and information, and uploads the data to the server. The server then preprocesses the received data, performs risk assessment using a machine learning model, and recommends safe trading assets to the user. Each component and its operation are described in detail below.

[0085] Terminal data collection and upload

[0086] The device periodically scans data such as the user's emails, chat messages, calendar events, and meeting minutes. For example, a Python script installed on the device periodically checks the mailbox to detect new emails. The device accesses the user's mailbox using the Microsoft Graph API to retrieve new emails. This collected data is stored as a temporary file on the device's local disk. Then, at a specified time or triggered by a specific event, the data is uploaded to a server using the HTTPS protocol.

[0087] Data preprocessing by the server

[0088] When the server receives data sent from a terminal, it first preprocesses the data. A Flask-based web server receives the POST request and saves the data in the appropriate folder. Preprocessing includes standardizing the text data (standardizing character codes), tokenizing it using natural language processing (dividing sentences into words), and removing commonly used words (removing frequently used words such as "no" and "ni"). For example, HTML tags can be removed from the contents of meeting minutes to format them as pure text.

[0089] Risk Assessment

[0090] Based on the preprocessed data, the server uses a machine learning model to perform risk assessment. A TensorFlow model is used to score the likelihood of insider trading. This score is evaluated on a scale of 0 to 100; for example, an email containing a discussion about a new product from X Corporation would be calculated as a risk score of 90. The machine learning model is trained based on past market manipulation data, allowing it to quickly and accurately assess risk on new data.

[0091] Risk score aggregation

[0092] The risk scores obtained for each data item are aggregated by the server. Higher scores are excluded from trading, and lower scores are listed. For example, if "Company X's risk score is 90" and "Company Y's risk score is 20," then "Company Y" will be included in the recommendation list.

[0093] Stock recommendations and notifications

[0094] Based on the aggregated risk scores, the server recommends safe trading assets to the user. The recommendation results are sent to the terminal, which notifies the user through a notification system. For example, a specific list of stocks may be displayed, such as "The following stocks are safe for you to buy and sell today: Y Co., Ltd." The terminal uses the notification function to provide information to the user in real time.

[0095] Examples of concrete examples and prompts

[0096] For example, if a user sends or receives an email with the content "We will discuss a new product at the next board meeting," this email is collected by the device and uploaded to the server. The server preprocesses this email and inputs it into a machine learning model, which may determine that the email poses a high insider trading risk. In this case, the risk score will be high and the related stocks will not be recommended. Instead, other low-risk stocks will be recommended to the user.

[0097] Below are some example prompts for generative AI models:

[0098] "We have detected an email containing information about a new product that the user would like to discuss at the next board meeting. Based on this information, we would like you to calculate an insider trading risk score and recommend safe stocks to trade."

[0099] In this way, the system of the present invention works in conjunction with various communication means to assist users in safely and efficiently trading stocks without unintentionally incurring legal risks.

[0100] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0101] Step 1:

[0102] Data collection

[0103] The device periodically collects data from the user's emails, chat messages, calendar events, meeting minutes, etc. For example, a Python script installed on the device uses the Microsoft Graph API to access the user's mailbox and retrieve new emails.

[0104] Input: Various user communication data (emails, chat messages, etc.)

[0105] Output: Raw data stored on the device

[0106] Step 2:

[0107] Data storage and trigger settings

[0108] The collected data is saved as a temporary file on the device's local disk. For example, meeting minutes are saved as "meeting_notes.txt." The device also uploads data at a specified time (e.g., midnight) or when a specific event (e.g., receiving an email) triggers the upload.

[0109] Input: Raw data stored on the device

[0110] Output: Prepared upload data

[0111] Step 3:

[0112] Data upload

[0113] The data is encrypted and uploaded to the server using the HTTPS protocol: the script on the device sends the collected data to the server using HTTP POST requests.

[0114] Input: Prepared upload data

[0115] Output: Data sent to the server

[0116] Step 4:

[0117] Data reception

[0118] The server receives the data sent from the device. For example, a Flask-based web server receives a POST request and saves the data in the appropriate folder (e.g., " / data / incoming").

[0119] Input: Data sent from the terminal

[0120] Output: Temporary data on the server

[0121] Step 5:

[0122] Data Preprocessing

[0123] The server performs preprocessing on the received data, including standardization, tokenization, and stop word removal. Specifically, it uses Python's Beautiful Soup library to remove HTML tags, and the NLTK library to tokenize each word and remove frequent words. For example, it removes HTML tags from the content of "meeting_notes.txt" and converts it into pure text.

[0124] Input: Temporary data on the server

[0125] Output: Preprocessed data

[0126] Step 6:

[0127] Risk Assessment

[0128] Based on the preprocessed data, the server uses machine learning models to perform risk assessments. For example, TensorFlow can be used to calculate a risk score that indicates the likelihood of insider trading. For example, minutes of a meeting about a new product might be assigned a high risk score (e.g., 90).

[0129] Input: Preprocessed data

[0130] Output: Risk score

[0131] Step 7:

[0132] Risk score aggregation

[0133] The server aggregates the risk scores obtained for each data. For example, if "Company A's risk score is 90" and "Company B's risk score is 20," then "Company B" will be included in the recommendation list.

[0134] Input: Risk Score

[0135] Output: Aggregated risk level

[0136] Step 8:

[0137] Stock Recommendations

[0138] The server recommends safe trading assets to users based on the aggregated risk level. Specifically, it generates a recommended list of assets with low risk levels.

[0139] Input: Aggregated risk level

[0140] Output: Recommended stock list

[0141] Step 9:

[0142] Notification of recommendation results

[0143] The server sends the recommendation results to the terminal, which then notifies the user. Specifically, the terminal's notification system is used to provide real-time information to the user. For example, it may notify the user that "This stock is safe to trade today: Co. B."

[0144] Input: Recommended stock list

[0145] Output: User notification

[0146] By following the steps above, users can safely trade stocks while minimizing the risk of insider trading.

[0147] (Application example 1)

[0148] 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."

[0149] Currently, it is extremely important to properly manage insider trading risks in stock trading. However, existing systems inadequately assess the risk of stocks traded by users, and manual confirmation work requires a great deal of effort. Furthermore, even when safe stocks are recommended, the procedures for immediate purchase are cumbersome, hindering smooth trading. Therefore, the present invention aims to provide a system that supports users in safely and efficiently trading stocks, enabling risk assessment and immediate purchase.

[0150] 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.

[0151] In this invention, the server includes means for evaluating a risk score associated with insider trading using a machine learning model based on the preprocessed data, means for aggregating multiple risk scores to determine a final risk level, and means for instantly purchasing safe stocks in cooperation with the user's electronic payment platform based on the aggregated risk level, thereby enabling the user to quickly purchase safe trading stocks.

[0152] A "terminal" is an electronic device used by a user, such as a smartphone or PC, that collects and notifies data.

[0153] A "server" is a computer system that receives, processes, and evaluates data sent from a terminal.

[0154] "Communication tools" are tools used by users, such as email, chat messages, and calendar events.

[0155] "Materials" refers to document data related to a transaction, such as a user's electronic documents or meeting minutes.

[0156] "Data collection" is the process by which the device periodically collects data from the user's communication tools and materials.

[0157] "Data preprocessing" is the process by which the server standardizes, tokenizes, removes stop words, etc., data sent from the terminal.

[0158] A "machine learning model" is a model that is trained based on past insider trading data and performs risk assessments on new data.

[0159] The "risk score" is a numerical representation of insider trading risk calculated using a machine learning model.

[0160] The "risk level" is an index that indicates the safety of a trading stock and is determined by aggregating multiple risk scores.

[0161] "Trading assets" are financial assets such as stocks and securities that users buy and sell.

[0162] "Recommendation" refers to the act of the server suggesting trading assets to a user based on risk level.

[0163] "Notification" is the process by which the terminal notifies the user of the recommendation results received from the server.

[0164] An "electronic payment platform" is an online payment system used to purchase and sell trading assets.

[0165] "Linkage" is the process by which the server communicates data with the electronic payment platform to realize the transaction.

[0166] The system of the present invention involves a terminal and a server working together to evaluate the risk of insider trading in stock trading and recommend safe stocks to users. The components of this system and their operation are described below.

[0167] Data collection

[0168] The device periodically collects data from the communication tools (emails, chat messages, calendar events, etc.) and documents used by the user. This collected data is temporarily stored on the device.

[0169] Uploading data

[0170] At a specified time or trigger, the collected data is consolidated and uploaded to the server by the device, and the uploaded data is securely transmitted to the server via the Internet.

[0171] Data Preprocessing

[0172] The server preprocesses the received data, which includes standardizing the text data (removing HTML tags, etc.), tokenizing (splitting words), removing stop words (removing non-important words), etc. Spacy is used as the natural language processing library.

[0173] Risk Assessment

[0174] Based on the preprocessed text data, the server uses a generative AI model (machine learning model) to calculate a risk score. This model is trained on past insider trading data and assesses the insider trading risk of new data.

[0175] Determining the risk level

[0176] The server aggregates multiple risk scores and determines a final risk level, which evaluates the overall insider trading risk for each stock.

[0177] Stock recommendation and electronic payment platform integration

[0178] The server recommends safe trading assets to users based on the aggregated risk level, and also provides a means for users to instantly purchase safe stocks in conjunction with their electronic payment platform, allowing users to quickly purchase the recommended stocks.

[0179] Notification of recommendation results

[0180] Finally, the server sends the recommendation results to the device, which then notifies the user via a smartphone application or a PC notification system.

[0181] Specific examples

[0182] For example, if a user sends or receives an email with the content "At the next meeting, we will discuss a new product from Company X," this email is collected by the device and uploaded to the server. The server preprocesses this email and inputs it into a machine learning model to determine that it poses a high risk of insider trading. In this case, the risk score will be high, and the related Company X's stock will be excluded as a non-recommended stock. Meanwhile, other low-risk stocks will be recommended to the user.

[0183] An example of a prompt sentence to input to the generative AI model is as follows:

[0184] At the next meeting, we will discuss Company X's new product.

[0185] Yu

[0186] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0187] Step 1:

[0188] The device collects data from the user's communication tools and documents. Input data includes emails, chat messages, calendar events, etc., and periodically scans them to extract important information and temporarily store it on the device. The output is a list of the collected data.

[0189] Step 2:

[0190] The device uploads the collected data to the server. Here, the data is encrypted and securely sent to the server at regular intervals or when a specific trigger (user-specified operation or detection of new data) is triggered. The input is the data stored on the device, and the output is the data received by the server.

[0191] Step 3:

[0192] The server preprocesses the received data. Specifically, it standardizes, tokenizes, and removes stop words. Spacy is used as a natural language processing library, and the input data is text data, and the output is preprocessed, cleaned text data.

[0193] Step 4:

[0194] The server evaluates a risk score associated with insider trading using a generative AI model based on the preprocessed data. The machine learning model is trained on historical insider trading data and takes the preprocessed text data as input. The output is a calculated insider trading risk score.

[0195] Step 5:

[0196] The server aggregates multiple risk scores and determines the final risk level. Here, the risk scores for each data are integrated to perform an overall risk assessment. The input is the individual risk scores, and the output is the integrated final risk level.

[0197] Step 6:

[0198] The server recommends safe trading assets to users based on their risk level. Depending on the risk level, stocks deemed safe are selected and a list is generated. The input is the final risk level, and the output is a list of recommended safe trading assets.

[0199] Step 7:

[0200] The server instantly purchases safe stocks based on the aggregated risk level in conjunction with the user's electronic payment platform. The server uses the user's account information and payment system to automate the purchase process for the recommended stocks. The input is a list of recommended trading assets, and the output is confirmation of the completion of the purchase process.

[0201] Step 8:

[0202] The terminal notifies the user of the recommendation results and purchase completion notification. Specifically, the recommended stocks and purchase results are notified to the user via a smartphone application or a PC notification system. The input is the notification content sent from the server, and the output is what is displayed to the user.

[0203] 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.

[0204] The system of the present invention operates in cooperation with a terminal and a server to support users in safely trading stocks. This system also incorporates an emotion engine that recognizes the user's emotions, and the results can be used to assess the risk of insider trading. Specific embodiments are described below.

[0205] Data collection and upload

[0206] The device collects data such as the user's emails, chat messages, voice data, calendar events, and meeting minutes. This collection process is automated and runs periodically. For example, the device accesses the user's mailbox every hour to retrieve new emails and voice messages. Calendar and meeting minutes data is also retrieved in the same way. This data is temporarily stored in the device's local storage and periodically uploaded to the server.

[0207] Data Preprocessing

[0208] When the server receives data sent from a terminal, it first preprocesses the data. Preprocessing includes cleaning the text data. Specifically, HTML tags are removed from emails and chat messages to convert them into pure text. Natural language processing is then used to tokenize the data and remove meaningless words (stop words). In the case of audio data, speech recognition technology is used to convert it into text.

[0209] Emotion analysis

[0210] The server uses the preprocessed text and voice data to perform emotion analysis using an emotion engine. The emotion engine recognizes the user's emotions (happiness, anger, sadness, etc.) from the text data. It also analyzes the tone and intonation of the voice data to recognize similar emotions. This emotion data is then incorporated into the risk assessment process.

[0211] Risk Assessment

[0212] Once preprocessing and sentiment analysis are complete, the server uses a machine learning model to perform risk assessment on the data. The model analyzes the preprocessed text data, audio data, and sentiment data together to calculate a risk score related to insider trading. This score indicates the risk level of each data.

[0213] Risk score aggregation

[0214] The risk scores calculated for each data item are aggregated by the server, and the final risk level is determined. Specifically, the risk scores obtained for each data item are aggregated using statistical methods to calculate the overall risk level for each stock. This allows an evaluation of whether a particular stock can be traded.

[0215] Stock recommendations and notifications

[0216] The server recommends safe stocks to the user based on the aggregated risk level. This involves the process of selecting and listing stocks with low risk scores. The selected stocks are recommended to the user and ultimately sent to the user's device. The user's device receives the recommendation results and notifies the user. The device displays a pop-up notification or email with a "list of stocks that are safe to buy and sell today," providing the user with information to trade safely. This notification allows the user to proceed with trading with peace of mind.

[0217] Specific examples

[0218] For example, if a user sends and receives an email stating, "At the next board meeting, we will discuss a new product from Company X," along with a voice message expressing joy, this data is collected by the device and uploaded to the server. The server preprocesses this email and voice data and uses an emotion engine to recognize that the user is expressing joy. The server then inputs this data into a machine learning model and determines that the insider trading risk is high (high risk score). In this case, shares of Company X, which has a high risk level, are not recommended to the user, and instead other low-risk stocks are recommended to the user.

[0219] This system allows users to trade stocks safely without worrying about the risk of insider trading, and reduces legal risks for the company as a whole. The addition of an emotion engine enables risk assessment that takes user emotions into account, resulting in more accurate recommendations. Each component of this system is built by combining existing technologies, making it a widely applicable and feasible solution.

[0220] The processing flow will be explained below.

[0221] Step 1:

[0222] The device collects data such as the user's emails, chat messages, voice data, calendar events, and meeting minutes. This collection process is automated, so for example, the device accesses the user's mailbox every 30 minutes to retrieve new emails and newly recorded voice data. Calendar information and meeting minutes data are also collected in the same way.

[0223] Step 2:

[0224] The device uploads the collected data to the server. The collected data is temporarily stored in the device's local storage and then transferred to the server based on a set schedule. At this time, the data is encrypted to ensure security.

[0225] Step 3:

[0226] The server preprocesses the data it receives. Preprocessing includes cleaning the text data. For example, HTML tags are removed from emails and chat messages to convert them into pure text. Natural language processing technology is then used to tokenize the data and remove meaningless words (stop words). In the case of audio data, speech recognition technology is used to convert it into text.

[0227] Step 4:

[0228] The server uses an emotion engine to analyze the user's emotions from the preprocessed text and voice data. The emotion engine uses, for example, natural language processing (NLP) techniques to recognize the user's emotions (happiness, anger, sadness, etc.) from the text data. It also analyzes the tone and intonation of the voice data to recognize similar emotions. The recognized emotion data is then incorporated into the subsequent risk assessment process.

[0229] Step 5:

[0230] The server uses the sentiment-analyzed data to perform risk assessment using a machine learning model. The machine learning model is a trained model that calculates a risk score related to insider trading. The model takes the preprocessed text data, audio data, and sentiment data as input and outputs a numerical risk level for each data.

[0231] Step 6:

[0232] The server aggregates multiple risk scores to determine the final risk level. Specifically, the risk scores obtained for each data point are aggregated using statistical methods such as weighted averaging to calculate an overall risk level for each stock. This is used to evaluate whether a particular stock can be traded.

[0233] Step 7:

[0234] The server recommends safe trading stocks to the user based on the risk level. Stocks with low risk scores are selected and listed. This recommendation list is sent from the server to the user's terminal.

[0235] Step 8:

[0236] The terminal notifies the user of the recommendation results sent from the server. Specifically, the terminal displays a "list of stocks that are safe to buy and sell today" via a pop-up notification or email, providing the user with information to trade safely. This notification allows the user to trade with peace of mind.

[0237] Example 2

[0238] 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."

[0239] The challenge is to provide a system that reduces the risk of insider trading and supports safe stock trading. Conventional systems do not take user sentiment or unstructured data into account, which can result in low accuracy in risk assessment and the possibility of incorrect trade recommendations. Furthermore, data preprocessing and risk assessment require a significant amount of time and effort, so automation and improved accuracy are required.

[0240] 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.

[0241] In this invention, the server includes means for preprocessing received data, means for analyzing user emotions based on the preprocessed data using an emotion recognition engine, and means for evaluating a risk score related to insider trading using a machine learning model based on the emotion-analyzed data. This enables highly accurate risk assessment that takes user emotions and unstructured data into consideration, and the recommendation of safe stocks to trade.

[0242] "Terminal" means a computing device used by a user, primarily a hardware device or software application for data collection, processing, and notification.

[0243] "Server" refers to a computer system that receives data sent from terminals and performs centralized processing such as pre-processing, sentiment analysis, risk assessment, and stock recommendations.

[0244] An "emotion recognition engine" is a machine learning model or algorithm that analyzes a user's emotions from text and voice data and identifies the type and intensity of those emotions.

[0245] A "machine learning model" is an algorithm that learns patterns and trends based on past data and makes predictions and classifications for new data, and is used particularly for risk assessment in the present invention.

[0246] A "risk score" is a numerical representation of the degree of risk associated with insider trading, and is an index calculated for each data point.

[0247] "Preprocessing" is the process of applying initial processing to received data, including text standardization, tokenization, stop word removal, and converting audio data to text.

[0248] "Communication Tools" means software or platforms that allow users to send and receive messages and make voice and video calls, such as email clients and chat applications.

[0249] An "information source" is a source of digital content that a user accesses, such as a mail server, chat application, or calendar system.

[0250] "Uploading" is the process by which a device sends collected data to a server, using a secure transfer protocol such as SFTP or HTTPS.

[0251] "Stock recommendation" is the process by which the server suggests safe trading stocks to users based on risk assessment.

[0252] "Notification" refers to the process by which the device notifies the user of the recommendation results sent from the server, and includes providing information via pop-up notifications or emails.

[0253] System Overview

[0254] The system of the present invention, in which a terminal and a server work in cooperation with each other, supports users in safely trading stocks. This system also incorporates an emotion engine that recognizes the user's emotions, and the results can be used to assess the risk of insider trading. Specific embodiments are described below.

[0255] Data collection by terminal

[0256] The device collects data such as the user's emails, chat messages, voice data, calendar events, and meeting minutes. This collection process is automated and runs periodically. For example, the device accesses the user's mailbox every hour to retrieve new emails and voice messages. This is done using the IMAP protocol. Additionally, calendar events are retrieved using the Google Calendar API. This data is temporarily stored in the device's local storage and then securely uploaded to the server using, for example, SFTP or HTTPS.

[0257] Data preprocessing by the server

[0258] When the server receives data sent from the device, it first preprocesses the data. HTML tags are removed from emails and chat messages and converted into pure text. This is done using Python's BeautifulSoup. Natural language processing techniques are then used to tokenize the text data and remove stop words. Voice data is converted into text using the Google Speech-to-Text API.

[0259] Emotion analysis by server

[0260] Based on the pre-processed text and audio data, the server performs sentiment analysis using an emotion engine. The text data is input into a sentiment analysis model (e.g., BERT) and an emotion label is output. The audio data is analyzed for tone and intonation to identify emotions. This emotion data is used in the subsequent risk assessment process.

[0261] Server Risk Assessment

[0262] After data preprocessing and sentiment analysis are completed, the server performs risk assessment using a machine learning model (e.g., XGBoost). It calculates an insider trading risk score based on the preprocessed text data, voice data, and sentiment data. This risk score indicates the risk level of each data.

[0263] Risk score aggregation and stock recommendations

[0264] The server uses statistical methods to compile the risk scores calculated for each data point and determine the overall risk level for each stock. Stocks that are deemed safe are sent from the server to the user's device as recommendation information. The device receives this recommendation and presents the user with a "list of stocks that are safe to buy and sell today" via a pop-up notification or email.

[0265] Specific examples

[0266] For example, if a user sends or receives an email stating, "At the next meeting, we will discuss Company A's new product," along with a voice message expressing their feelings of joy, this data is collected by the device. The device then uploads this email and voice data to the server. The server then preprocesses the data by removing HTML tags and converting it into text. The emotion engine then recognizes the user's emotion of joy. This data is then input into a machine learning model, which evaluates the risk of insider trading as high. In this case, Company A's stock, which has a high risk level, is not recommended, and instead other low-risk stocks are recommended to the user.

[0267] Prompt Sentence Examples

[0268] "Please explain in detail how the system preprocesses data collected by the device, including emails, chat messages, voice data, calendar events, and meeting minutes, and then uses an emotion engine to perform risk assessments. This ensures secure transactions."

[0269] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0270] Step 1: Data collection

[0271] The device periodically collects data such as the user's emails, chat messages, voice data, calendar events, and meeting minutes. For example, the device accesses the mail server every hour using the IMAP protocol to download new emails. It also retrieves calendar events using the Google Calendar API and saves them in local storage.

[0272] Input: Mail server, chat app, voice data source, calendar system

[0273] Output: New emails, chat messages, audio files, and calendar events stored in local storage.

[0274] Step 2: Upload data

[0275] The device uploads the collected data to a server at regular intervals, and to ensure security, the data is transferred using SFTP or HTTPS protocols.

[0276] Input: Data in local storage (emails, chat messages, audio files, calendar events)

[0277] Output: Report that data upload to server is complete

[0278] Step 3: Data Preprocessing

[0279] The server preprocesses the data received from the device. First, it uses Python's BeautifulSoup to remove HTML tags from emails and chat messages and convert them into pure text. Next, it tokenizes the text data using natural language processing techniques and removes stop words. For audio data, it converts it to text using the Google Speech-to-Text API.

[0280] Input: Uploaded emails, chat messages, audio files, calendar events

[0281] Output: Preprocessed clean text data and converted audio data

[0282] Step 4: Sentiment Analysis

[0283] The server runs an emotion engine using the preprocessed text and audio data. The emotion engine uses an emotion analysis model such as the BERT model to recognize user emotions (e.g., joy, anger, sadness) from the text data. It also analyzes tone and intonation from the audio data to recognize similar emotions.

[0284] Input: Preprocessed text data, audio data

[0285] Output: Emotional information tagged to text and audio data

[0286] Step 5: Risk assessment

[0287] The server calculates an insider trading risk score using a machine learning model (e.g., XGBoost) based on the pre-processed and sentiment-analyzed data. It integrates the pre-processed text data, voice data, and sentiment data to evaluate the risk level of each.

[0288] Input: Sentiment-analyzed text data, audio data

[0289] Output: Risk score for each data point

[0290] Step 6: Aggregate risk scores

[0291] The server then uses statistical methods to aggregate the calculated risk scores to determine the overall risk level for each security. This process involves evaluating the overall risk scores for each data point and identifying the security deemed to be the least risky.

[0292] Input: Individual Risk Score

[0293] Output: Overall risk level for each stock

[0294] Step 7: Stock Recommendation and Notification

[0295] The server recommends safe stocks to the user based on the aggregated risk level. The list of selected stocks is sent to the terminal. The terminal notifies the user via a pop-up notification or email of the "list of stocks that are safe to buy and sell today."

[0296] Input: Aggregated risk level

[0297] Output: Recommended stock notification to user device

[0298] In this way, the system of the present invention helps users to trade stocks safely.

[0299] (Application example 2)

[0300] 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."

[0301] In stock trading, there is a risk of insider trading and inappropriate trading due to users making decisions based on their emotions. Current systems perform risk assessment without taking users' emotions into account, which limits their accuracy. In addition, they lack protection against users making large transactions while in an emotional state. Therefore, there is a need for a system that performs risk assessment taking users' emotions into account and supports safe and appropriate trading.

[0302] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for converting voice data into text, means for recognizing emotions from the text data and voice data and performing risk assessment based on the emotions, and means for suspending a transaction and requesting confirmation in a high-risk situation. This enables highly accurate risk assessment that takes the user's emotions into consideration, and supports safe and appropriate transactions.

[0303] A "terminal" is a device used by a user that collects data and uploads it to a server.

[0304] The "server" is a central processing unit that receives data sent from the terminals and performs pre-processing and risk assessment.

[0305] "Communication tools" are communication means that users use on a daily basis, such as e-mail, chat applications, and voice message applications.

[0306] "Data preprocessing" refers to data processing in which the server cleans the data it receives, tokenizes it, removes stop words, and so on.

[0307] A "machine learning model" is an algorithm that is trained based on past data and used to assess risk.

[0308] "Insider trading" is the illegal act of trading stocks using inside information.

[0309] A "risk score" is a numerical representation of the risk level of a transaction, calculated using a machine learning model.

[0310] "Emotion recognition" means analyzing emotions (joy, anger, sadness, etc.) from a user's text data or voice data.

[0311] A "high-risk situation" is one in which the risk score is high and caution or restrictions on transactions are required.

[0312] "Trading products" are items such as financial products and stocks that are the subject of trading by users.

[0313] The system of the present invention evaluates insider trading risks while taking into account user sentiment and supports safe trading. Specifically, the system is composed of a terminal, a server, and a machine learning model.

[0314] Data collection and upload

[0315] The device automatically collects data from the communication tools and materials used by the user. The collected data includes emails, chat messages, voice data, calendar events, and meeting minutes. This process is performed periodically; for example, the device accesses the user's mailbox every hour to retrieve new emails. It also collects voice recordings of voice data. This data is temporarily stored in the device's local storage and periodically uploaded to the server.

[0316] Data Preprocessing

[0317] When the server receives the data sent from the device, it first preprocesses the data. This includes cleaning and standardizing the text data, tokenizing it, and removing stop words. In the case of audio data, it is converted into text using speech recognition technology. For this purpose, it is effective to use the Python SpeechRecognition library.

[0318] Emotion analysis

[0319] The server performs sentiment analysis using an emotion engine based on the preprocessed text and audio data. The emotion engine recognizes the user's emotions (happiness, anger, sadness, etc.) from the text and audio data. For text data, it uses the TextBlob library, and for audio data, it analyzes the tone and intonation of the voice.

[0320] Risk Assessment

[0321] Once preprocessing and sentiment analysis are complete, the server uses the data to perform risk assessment using a machine learning model. This model is trained on historical insider trading data and sentiment data. The machine learning model uses the scikit-learn library and applies the logistic regression algorithm.

[0322] Risk score aggregation and transaction restrictions

[0323] The risk scores calculated for each data item are aggregated by the server to determine the final risk level. If the risk level is high, the server sends a notification to the terminal requesting a pause in the transaction and requesting the user to reconfirm. This prevents users from making high-risk transactions based on emotional judgment.

[0324] Stock recommendations and notifications

[0325] The server recommends safe trading products to users based on the final risk level. This includes the process of selecting and listing low-risk trading products. The selected products are sent to the user's device and displayed via pop-up notification or email. This allows users to proceed with trading with peace of mind.

[0326] Specific examples

[0327] For example, if a user sends a voice message saying, "I want to make a big purchase right now!", the device collects this and sends it to the server, which converts the voice data into text and performs sentiment analysis. If the sentiment score is high and the transaction is deemed high risk, the server will pause the transaction and send a notification to the user requesting confirmation.

[0328] Prompt Sentence Examples

[0329] "Analyze voice messages from users expressing their desire to make a large purchase, assess the risk of electronic payment services based on that sentiment, and pause the transaction if necessary."

[0330] This allows users to avoid emotional decisions and conduct safe and appropriate transactions.

[0331] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0332] Step 1:

[0333] The device collects data from the communication tools and materials used by the user. The collection process is automated and runs periodically. Specifically, the device accesses the user's mailbox and chat application every hour to retrieve new emails and messages. The device also records the user's voice messages, which are included in the collected data. The input is the communication tool and voice data, and the output is the collected data.

[0334] Step 2:

[0335] The device uploads the collected data to the server. This is a periodic process in which data temporarily stored in the device's local storage is transferred to the server. The input is the collected data, and the output is the data transferred to the server. Specifically, the data is sent to the server using an HTTP request.

[0336] Step 3:

[0337] The server preprocesses the received data. Preprocessing includes cleaning and standardizing the text data, tokenizing it, and removing stop words. In the case of audio data, it is first converted into text using speech recognition technology. The input is the raw data uploaded to the server, and the output is the cleaned text data and the audio data converted into text. Specific operations use the SpeechRecognition and TextBlob libraries.

[0338] Step 4:

[0339] The server performs sentiment analysis based on the preprocessed text and audio data. The sentiment engine recognizes the user's sentiment from the text and audio data. The input is the preprocessed data, and the output is an emotion score. Specifically, it uses the TextBlob library to calculate the sentiment score for the text and analyzes the tone and intonation of the audio data.

[0340] Step 5:

[0341] The server performs risk assessment using a machine learning model based on the sentiment data and preprocessed data. This model is trained based on past insider trading data and sentiment data. The input is sentiment data and preprocessed data, and the output is a risk score. Specifically, it uses the scikit-learn library and applies the Logistic Regression algorithm.

[0342] Step 6:

[0343] The server aggregates the risk scores calculated for each data item and determines the final risk level. The input is multiple risk scores, and the output is the final risk level. Specifically, the scores are aggregated using a statistical aggregation method.

[0344] Step 7:

[0345] The server recommends safe trading products to the user based on the risk level. If the risk level is high, the server suspends the transaction and sends a notification to the terminal requesting confirmation. The input is the final risk level, and the output is a list of recommended trading products and a notification. Specific operations include notifying the user via a pop-up notification or email.

[0346] Step 8:

[0347] The terminal notifies the user of the recommendation results sent from the server. The input is the recommendation results and notification request from the server, and the output is a notification message to the user. Specifically, the terminal displays a "list of trading products that are safe to buy and sell today" via a pop-up notification or email.

[0348] This series of steps allows users to avoid the risks of emotional decisions and conduct safe and appropriate transactions.

[0349] 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.

[0350] 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.

[0351] 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.

[0352] [Second embodiment]

[0353] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0354] 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.

[0355] 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).

[0356] 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.

[0357] 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.

[0358] 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).

[0359] 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.

[0360] 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.

[0361] 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.

[0362] 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.

[0363] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0364] 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."

[0365] The system of the present invention, in which terminals and a server work in cooperation with each other, supports users in safely trading stocks. This system operates in the following manner: the user's terminal collects data from communication tools and documents, uploads it to the server, the server preprocesses the data, performs risk assessment, and finally recommends safe stocks to the user.

[0366] Data collection and upload

[0367] The device periodically collects data such as the user's emails, chat messages, calendar events, meeting minutes, etc. This collected data is temporarily stored on the device and then uploaded to the server at a specified time or trigger.

[0368] Data Preprocessing

[0369] When the server receives data sent from the terminal, it first preprocesses the data. Preprocessing includes standardizing the text data, tokenizing it using natural language processing, removing stop words, etc. For example, HTML tags are removed from the email body and the email is formatted as pure text.

[0370] Risk Assessment

[0371] Based on the pre-processed text data, the server performs risk assessment using a pre-trained machine learning model, which is trained on historical insider trading data and calculates an insider trading risk score for new data.

[0372] Risk score aggregation

[0373] The risk scores calculated for each data point are aggregated by the server to determine a final risk level, which is then associated with each individual stock and provides an assessment of the stock the user plans to trade.

[0374] Stock recommendations and notifications

[0375] The server recommends safe stocks to the user based on the aggregated risk level. The results are sent to the terminal and notified to the user. For example, a list of specific stocks is displayed in the form of "These are the stocks that you can safely buy and sell today."

[0376] Specific examples

[0377] For example, if a user sends or receives an email stating, "At the next board meeting, we will discuss a new product from Company X," this email is collected by the device and uploaded to the server. The server preprocesses this email and inputs it into a machine learning model to determine that it is likely to pose a high insider trading risk. In this case, the risk score will be high, and the related Company X stock will be excluded as a non-recommended stock. Meanwhile, other low-risk stocks will be recommended to the user.

[0378] This system allows users to trade stocks safely without being aware of the risk of insider trading, and also reduces legal risks for the company as a whole. Each component of this system is built by combining existing technologies, and is widely available as a feasible means.

[0379] The processing flow will be explained below.

[0380] Step 1:

[0381] The device collects data such as the user's email, calendar, meeting minutes, and chat messages. This collection process is automated and runs periodically. Specifically, the device accesses the user's mailbox at specified intervals to retrieve new emails. Calendar and meeting minutes data is also collected in the same way.

[0382] Step 2:

[0383] The device uploads the collected data to the server. The collected data is temporarily stored in local storage and then transferred to the server according to the upload schedule. At this time, the data is encrypted for security reasons.

[0384] Step 3:

[0385] The server pre-processes the data it receives. This includes cleaning the text data. Specifically, HTML tags are removed from emails and chat messages, converting them to pure text, and then tokenizing them using natural language processing to remove meaningless words (stop words).

[0386] Step 4:

[0387] The server inputs the pre-processed text data into a machine learning model, a pre-trained risk assessment model, which outputs a score indicating the degree of risk associated with insider trading for each piece of data.

[0388] Step 5:

[0389] The server aggregates multiple risk scores to determine the final risk level. Specifically, it aggregates the risk scores obtained for each data point using statistical methods to calculate an overall risk level for each stock. This allows it to evaluate whether a particular stock can be traded.

[0390] Step 6:

[0391] The server recommends safe trading stocks to users based on their risk level, which includes a process of selecting and listing stocks with low risk scores. The selected stocks are then recommended to users.

[0392] Step 7:

[0393] The terminal notifies the user of the recommendation results sent from the server. Specifically, the terminal displays a "list of stocks that are safe to buy and sell today" via a pop-up notification or email, providing the user with information to trade safely. This notification allows the user to proceed with trading with peace of mind.

[0394] Example 1

[0395] 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."

[0396] There is a need to provide an effective support system to prevent users from unintentionally causing legal problems in stock trading, which involves risks such as market manipulation and insider trading. This system must appropriately collect data from the communication methods and information used by users, perform risk assessments based on that data, and recommend safe trading assets.

[0397] 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.

[0398] In this invention, the server includes a means for preprocessing the received data, a means for evaluating a risk score associated with market manipulation using a machine learning model based on the preprocessed data, and a means for aggregating multiple risk scores to determine a final risk level, thereby enabling the recommendation of safe trading assets for users to trade stocks with peace of mind.

[0399] A "terminal" is an electronic device used by a user, which has a means of communication and the ability to collect information and upload it to a server.

[0400] "Data collection" is the process by which the terminal obtains the necessary data from the user's communication means and information.

[0401] A "server" is a central computer system that processes data received from terminals and provides functions such as risk assessment and asset recommendations.

[0402] "Data preprocessing" is the process by which the server converts the raw data it receives into an analyzable format, and specifically includes text standardization, tokenization, and removal of commonly used words.

[0403] A "machine learning model" refers to an algorithm or mathematical model that learns patterns from data and makes predictions or classifications for new data.

[0404] A "risk score" is a numerical assessment of the risk associated with market manipulation and insider trading, assessed using a machine learning model.

[0405] The "risk level" is a comprehensive index that aggregates multiple risk scores and indicates the overall degree of risk.

[0406] "Safe trading assets" refer to assets that are deemed to have a low risk level and that users can trade with confidence.

[0407] "Recommendation" refers to the act of the server recommending a transaction to the user based on the results of the risk assessment.

[0408] "Notification" is the process by which the terminal notifies the user of the recommendation results received from the server.

[0409] MODE FOR CARRYING OUT THE INVENTION

[0410] The system of the present invention aims to help users safely trade stocks, and involves the cooperation of a terminal and a server. Specifically, the terminal collects data from the user's various communication methods and information, and uploads the data to the server. The server then preprocesses the received data, performs risk assessment using a machine learning model, and recommends safe trading assets to the user. Each component and its operation are described in detail below.

[0411] Terminal data collection and upload

[0412] The device periodically scans data such as the user's emails, chat messages, calendar events, and meeting minutes. For example, a Python script installed on the device periodically checks the mailbox to detect new emails. The device accesses the user's mailbox using the Microsoft Graph API to retrieve new emails. This collected data is stored as a temporary file on the device's local disk. Then, at a specified time or triggered by a specific event, the data is uploaded to a server using the HTTPS protocol.

[0413] Data preprocessing by the server

[0414] When the server receives data sent from a terminal, it first preprocesses the data. A Flask-based web server receives the POST request and saves the data in the appropriate folder. Preprocessing includes standardizing the text data (standardizing character codes), tokenizing it using natural language processing (dividing sentences into words), and removing commonly used words (removing frequently used words such as "no" and "ni"). For example, HTML tags can be removed from the contents of meeting minutes to format them as pure text.

[0415] Risk Assessment

[0416] Based on the preprocessed data, the server uses a machine learning model to perform risk assessment. A TensorFlow model is used to score the likelihood of insider trading. This score is evaluated on a scale of 0 to 100; for example, an email containing a discussion about a new product from X Corporation would be calculated as a risk score of 90. The machine learning model is trained based on past market manipulation data, allowing it to quickly and accurately assess risk on new data.

[0417] Risk score aggregation

[0418] The risk scores obtained for each data item are aggregated by the server. Higher scores are excluded from trading, and lower scores are listed. For example, if "Company X's risk score is 90" and "Company Y's risk score is 20," then "Company Y" will be included in the recommendation list.

[0419] Stock recommendations and notifications

[0420] Based on the aggregated risk scores, the server recommends safe trading assets to the user. The recommendation results are sent to the terminal, which notifies the user through a notification system. For example, a specific list of stocks may be displayed, such as "The following stocks are safe for you to buy and sell today: Y Co., Ltd." The terminal uses the notification function to provide information to the user in real time.

[0421] Examples of concrete examples and prompts

[0422] For example, if a user sends or receives an email with the content "We will discuss a new product at the next board meeting," this email is collected by the device and uploaded to the server. The server preprocesses this email and inputs it into a machine learning model, which may determine that the email poses a high insider trading risk. In this case, the risk score will be high and the related stocks will not be recommended. Instead, other low-risk stocks will be recommended to the user.

[0423] Below are some example prompts for generative AI models:

[0424] "We have detected an email containing information about a new product that the user would like to discuss at the next board meeting. Based on this information, we would like you to calculate an insider trading risk score and recommend safe stocks to trade."

[0425] In this way, the system of the present invention works in conjunction with various communication means to assist users in safely and efficiently trading stocks without unintentionally incurring legal risks.

[0426] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0427] Step 1:

[0428] Data collection

[0429] The device periodically collects data from the user's emails, chat messages, calendar events, meeting minutes, etc. For example, a Python script installed on the device uses the Microsoft Graph API to access the user's mailbox and retrieve new emails.

[0430] Input: Various user communication data (emails, chat messages, etc.)

[0431] Output: Raw data stored on the device

[0432] Step 2:

[0433] Data storage and trigger settings

[0434] The collected data is saved as a temporary file on the device's local disk. For example, meeting minutes are saved as "meeting_notes.txt." The device also uploads data at a specified time (e.g., midnight) or when a specific event (e.g., receiving an email) triggers the upload.

[0435] Input: Raw data stored on the device

[0436] Output: Prepared upload data

[0437] Step 3:

[0438] Data upload

[0439] The data is encrypted and uploaded to the server using the HTTPS protocol: the script on the device sends the collected data to the server using HTTP POST requests.

[0440] Input: Prepared upload data

[0441] Output: Data sent to the server

[0442] Step 4:

[0443] Data reception

[0444] The server receives the data sent from the device. For example, a Flask-based web server receives a POST request and saves the data in the appropriate folder (e.g., " / data / incoming").

[0445] Input: Data sent from the terminal

[0446] Output: Temporary data on the server

[0447] Step 5:

[0448] Data Preprocessing

[0449] The server performs preprocessing on the received data, including standardization, tokenization, and stop word removal. Specifically, it uses Python's Beautiful Soup library to remove HTML tags, and the NLTK library to tokenize each word and remove frequent words. For example, it removes HTML tags from the content of "meeting_notes.txt" and converts it into pure text.

[0450] Input: Temporary data on the server

[0451] Output: Preprocessed data

[0452] Step 6:

[0453] Risk Assessment

[0454] Based on the preprocessed data, the server uses machine learning models to perform risk assessments. For example, TensorFlow can be used to calculate a risk score that indicates the likelihood of insider trading. For example, minutes of a meeting about a new product might be assigned a high risk score (e.g., 90).

[0455] Input: Preprocessed data

[0456] Output: Risk score

[0457] Step 7:

[0458] Risk score aggregation

[0459] The server aggregates the risk scores obtained for each data. For example, if "Company A's risk score is 90" and "Company B's risk score is 20," then "Company B" will be included in the recommendation list.

[0460] Input: Risk Score

[0461] Output: Aggregated risk level

[0462] Step 8:

[0463] Stock Recommendations

[0464] The server recommends safe trading assets to users based on the aggregated risk level. Specifically, it generates a recommended list of assets with low risk levels.

[0465] Input: Aggregated risk level

[0466] Output: Recommended stock list

[0467] Step 9:

[0468] Notification of recommendation results

[0469] The server sends the recommendation results to the terminal, which then notifies the user. Specifically, the terminal's notification system is used to provide real-time information to the user. For example, it may notify the user that "This stock is safe to trade today: Co. B."

[0470] Input: Recommended stock list

[0471] Output: User notification

[0472] By following the steps above, users can safely trade stocks while minimizing the risk of insider trading.

[0473] (Application example 1)

[0474] 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."

[0475] Currently, it is extremely important to properly manage insider trading risks in stock trading. However, existing systems inadequately assess the risk of stocks traded by users, and manual confirmation work requires a great deal of effort. Furthermore, even when safe stocks are recommended, the procedures for immediate purchase are cumbersome, hindering smooth trading. Therefore, the present invention aims to provide a system that supports users in safely and efficiently trading stocks, enabling risk assessment and immediate purchase.

[0476] 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.

[0477] In this invention, the server includes means for evaluating a risk score associated with insider trading using a machine learning model based on the preprocessed data, means for aggregating multiple risk scores to determine a final risk level, and means for instantly purchasing safe stocks in cooperation with the user's electronic payment platform based on the aggregated risk level, thereby enabling the user to quickly purchase safe trading stocks.

[0478] A "terminal" is an electronic device used by a user, such as a smartphone or PC, that collects and notifies data.

[0479] A "server" is a computer system that receives, processes, and evaluates data sent from a terminal.

[0480] "Communication tools" are tools used by users, such as email, chat messages, and calendar events.

[0481] "Materials" refers to document data related to a transaction, such as a user's electronic documents or meeting minutes.

[0482] "Data collection" is the process by which the device periodically collects data from the user's communication tools and materials.

[0483] "Data preprocessing" is the process by which the server standardizes, tokenizes, removes stop words, etc., data sent from the terminal.

[0484] A "machine learning model" is a model that is trained based on past insider trading data and performs risk assessments on new data.

[0485] The "risk score" is a numerical representation of insider trading risk calculated using a machine learning model.

[0486] The "risk level" is an index that indicates the safety of a trading stock and is determined by aggregating multiple risk scores.

[0487] "Trading assets" are financial assets such as stocks and securities that users buy and sell.

[0488] "Recommendation" refers to the act of the server suggesting trading assets to a user based on risk level.

[0489] "Notification" is the process by which the terminal notifies the user of the recommendation results received from the server.

[0490] An "electronic payment platform" is an online payment system used to purchase and sell trading assets.

[0491] "Linkage" is the process by which the server communicates data with the electronic payment platform to realize the transaction.

[0492] The system of the present invention involves a terminal and a server working together to evaluate the risk of insider trading in stock trading and recommend safe stocks to users. The components of this system and their operation are described below.

[0493] Data collection

[0494] The device periodically collects data from the communication tools (emails, chat messages, calendar events, etc.) and documents used by the user. This collected data is temporarily stored on the device.

[0495] Uploading data

[0496] At a specified time or trigger, the collected data is consolidated and uploaded to the server by the device, and the uploaded data is securely transmitted to the server via the Internet.

[0497] Data Preprocessing

[0498] The server preprocesses the received data, which includes standardizing the text data (removing HTML tags, etc.), tokenizing (splitting words), removing stop words (removing non-important words), etc. Spacy is used as the natural language processing library.

[0499] Risk Assessment

[0500] Based on the preprocessed text data, the server uses a generative AI model (machine learning model) to calculate a risk score. This model is trained on past insider trading data and assesses the insider trading risk of new data.

[0501] Determining the risk level

[0502] The server aggregates multiple risk scores and determines a final risk level, which evaluates the overall insider trading risk for each stock.

[0503] Stock recommendation and electronic payment platform integration

[0504] The server recommends safe trading assets to users based on the aggregated risk level, and also provides a means for users to instantly purchase safe stocks in conjunction with their electronic payment platform, allowing users to quickly purchase the recommended stocks.

[0505] Notification of recommendation results

[0506] Finally, the server sends the recommendation results to the device, which then notifies the user via a smartphone application or a PC notification system.

[0507] Specific examples

[0508] For example, if a user sends or receives an email with the content "At the next meeting, we will discuss a new product from Company X," this email is collected by the device and uploaded to the server. The server preprocesses this email and inputs it into a machine learning model to determine that it poses a high risk of insider trading. In this case, the risk score will be high, and the related Company X's stock will be excluded as a non-recommended stock. Meanwhile, other low-risk stocks will be recommended to the user.

[0509] An example of a prompt sentence to input to the generative AI model is as follows:

[0510] At the next meeting, we will discuss Company X's new product.

[0511] Yu

[0512] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0513] Step 1:

[0514] The device collects data from the user's communication tools and documents. Input data includes emails, chat messages, calendar events, etc., and periodically scans them to extract important information and temporarily store it on the device. The output is a list of the collected data.

[0515] Step 2:

[0516] The device uploads the collected data to the server. Here, the data is encrypted and securely sent to the server at regular intervals or when a specific trigger (user-specified operation or detection of new data) is triggered. The input is the data stored on the device, and the output is the data received by the server.

[0517] Step 3:

[0518] The server preprocesses the received data. Specifically, it standardizes, tokenizes, and removes stop words. Spacy is used as a natural language processing library, and the input data is text data, and the output is preprocessed, cleaned text data.

[0519] Step 4:

[0520] The server evaluates a risk score associated with insider trading using a generative AI model based on the preprocessed data. The machine learning model is trained on historical insider trading data and takes the preprocessed text data as input. The output is a calculated insider trading risk score.

[0521] Step 5:

[0522] The server aggregates multiple risk scores and determines the final risk level. Here, the risk scores for each data are integrated to perform an overall risk assessment. The input is the individual risk scores, and the output is the integrated final risk level.

[0523] Step 6:

[0524] The server recommends safe trading assets to users based on their risk level. Depending on the risk level, stocks deemed safe are selected and a list is generated. The input is the final risk level, and the output is a list of recommended safe trading assets.

[0525] Step 7:

[0526] The server instantly purchases safe stocks based on the aggregated risk level in conjunction with the user's electronic payment platform. The server uses the user's account information and payment system to automate the purchase process for the recommended stocks. The input is a list of recommended trading assets, and the output is confirmation of the completion of the purchase process.

[0527] Step 8:

[0528] The terminal notifies the user of the recommendation results and purchase completion notification. Specifically, the recommended stocks and purchase results are notified to the user via a smartphone application or a PC notification system. The input is the notification content sent from the server, and the output is what is displayed to the user.

[0529] 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.

[0530] The system of the present invention operates in cooperation with a terminal and a server to support users in safely trading stocks. This system also incorporates an emotion engine that recognizes the user's emotions, and the results can be used to assess the risk of insider trading. Specific embodiments are described below.

[0531] Data collection and upload

[0532] The device collects data such as the user's emails, chat messages, voice data, calendar events, and meeting minutes. This collection process is automated and runs periodically. For example, the device accesses the user's mailbox every hour to retrieve new emails and voice messages. Calendar and meeting minutes data is also retrieved in the same way. This data is temporarily stored in the device's local storage and periodically uploaded to the server.

[0533] Data Preprocessing

[0534] When the server receives data sent from a terminal, it first preprocesses the data. Preprocessing includes cleaning the text data. Specifically, HTML tags are removed from emails and chat messages to convert them into pure text. Natural language processing is then used to tokenize the data and remove meaningless words (stop words). In the case of audio data, speech recognition technology is used to convert it into text.

[0535] Emotion analysis

[0536] The server uses the preprocessed text and voice data to perform emotion analysis using an emotion engine. The emotion engine recognizes the user's emotions (happiness, anger, sadness, etc.) from the text data. It also analyzes the tone and intonation of the voice data to recognize similar emotions. This emotion data is then incorporated into the risk assessment process.

[0537] Risk Assessment

[0538] Once preprocessing and sentiment analysis are complete, the server uses a machine learning model to perform risk assessment on the data. The model analyzes the preprocessed text data, audio data, and sentiment data together to calculate a risk score related to insider trading. This score indicates the risk level of each data.

[0539] Risk score aggregation

[0540] The risk scores calculated for each data item are aggregated by the server, and the final risk level is determined. Specifically, the risk scores obtained for each data item are aggregated using statistical methods to calculate the overall risk level for each stock. This allows an evaluation of whether a particular stock can be traded.

[0541] Stock recommendations and notifications

[0542] The server recommends safe stocks to the user based on the aggregated risk level. This involves the process of selecting and listing stocks with low risk scores. The selected stocks are recommended to the user and ultimately sent to the user's device. The user's device receives the recommendation results and notifies the user. The device displays a pop-up notification or email with a "list of stocks that are safe to buy and sell today," providing the user with information to trade safely. This notification allows the user to proceed with trading with peace of mind.

[0543] Specific examples

[0544] For example, if a user sends and receives an email stating, "At the next board meeting, we will discuss a new product from Company X," along with a voice message expressing joy, this data is collected by the device and uploaded to the server. The server preprocesses this email and voice data and uses an emotion engine to recognize that the user is expressing joy. The server then inputs this data into a machine learning model and determines that the insider trading risk is high (high risk score). In this case, shares of Company X, which has a high risk level, are not recommended to the user, and instead other low-risk stocks are recommended to the user.

[0545] This system allows users to trade stocks safely without worrying about the risk of insider trading, and reduces legal risks for the company as a whole. The addition of an emotion engine enables risk assessment that takes user emotions into account, resulting in more accurate recommendations. Each component of this system is built by combining existing technologies, making it a widely applicable and feasible solution.

[0546] The processing flow will be explained below.

[0547] Step 1:

[0548] The device collects data such as the user's emails, chat messages, voice data, calendar events, and meeting minutes. This collection process is automated, so for example, the device accesses the user's mailbox every 30 minutes to retrieve new emails and newly recorded voice data. Calendar information and meeting minutes data are also collected in the same way.

[0549] Step 2:

[0550] The device uploads the collected data to the server. The collected data is temporarily stored in the device's local storage and then transferred to the server based on a set schedule. At this time, the data is encrypted to ensure security.

[0551] Step 3:

[0552] The server preprocesses the data it receives. Preprocessing includes cleaning the text data. For example, HTML tags are removed from emails and chat messages to convert them into pure text. Natural language processing technology is then used to tokenize the data and remove meaningless words (stop words). In the case of audio data, speech recognition technology is used to convert it into text.

[0553] Step 4:

[0554] The server uses an emotion engine to analyze the user's emotions from the preprocessed text and voice data. The emotion engine uses, for example, natural language processing (NLP) techniques to recognize the user's emotions (happiness, anger, sadness, etc.) from the text data. It also analyzes the tone and intonation of the voice data to recognize similar emotions. The recognized emotion data is then incorporated into the subsequent risk assessment process.

[0555] Step 5:

[0556] The server uses the sentiment-analyzed data to perform risk assessment using a machine learning model. The machine learning model is a trained model that calculates a risk score related to insider trading. The model takes the preprocessed text data, audio data, and sentiment data as input and outputs a numerical risk level for each data.

[0557] Step 6:

[0558] The server aggregates multiple risk scores to determine the final risk level. Specifically, the risk scores obtained for each data point are aggregated using statistical methods such as weighted averaging to calculate an overall risk level for each stock. This is used to evaluate whether a particular stock can be traded.

[0559] Step 7:

[0560] The server recommends safe trading stocks to the user based on the risk level. Stocks with low risk scores are selected and listed. This recommendation list is sent from the server to the user's terminal.

[0561] Step 8:

[0562] The terminal notifies the user of the recommendation results sent from the server. Specifically, the terminal displays a "list of stocks that are safe to buy and sell today" via a pop-up notification or email, providing the user with information to trade safely. This notification allows the user to trade with peace of mind.

[0563] Example 2

[0564] 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."

[0565] The challenge is to provide a system that reduces the risk of insider trading and supports safe stock trading. Conventional systems do not take user sentiment or unstructured data into account, which can result in low accuracy in risk assessment and the possibility of incorrect trade recommendations. Furthermore, data preprocessing and risk assessment require a significant amount of time and effort, so automation and improved accuracy are required.

[0566] 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.

[0567] In this invention, the server includes means for preprocessing received data, means for analyzing user emotions based on the preprocessed data using an emotion recognition engine, and means for evaluating a risk score related to insider trading using a machine learning model based on the emotion-analyzed data. This enables highly accurate risk assessment that takes user emotions and unstructured data into consideration, and the recommendation of safe stocks to trade.

[0568] "Terminal" means a computing device used by a user, primarily a hardware device or software application for data collection, processing, and notification.

[0569] "Server" refers to a computer system that receives data sent from terminals and performs centralized processing such as pre-processing, sentiment analysis, risk assessment, and stock recommendations.

[0570] An "emotion recognition engine" is a machine learning model or algorithm that analyzes a user's emotions from text and voice data and identifies the type and intensity of those emotions.

[0571] A "machine learning model" is an algorithm that learns patterns and trends based on past data and makes predictions and classifications for new data, and is used particularly for risk assessment in the present invention.

[0572] A "risk score" is a numerical representation of the degree of risk associated with insider trading, and is an index calculated for each data point.

[0573] "Preprocessing" is the process of applying initial processing to received data, including text standardization, tokenization, stop word removal, and converting audio data to text.

[0574] "Communication Tools" means software or platforms that allow users to send and receive messages and make voice and video calls, such as email clients and chat applications.

[0575] An "information source" is a source of digital content that a user accesses, such as a mail server, chat application, or calendar system.

[0576] "Uploading" is the process by which a device sends collected data to a server, using a secure transfer protocol such as SFTP or HTTPS.

[0577] "Stock recommendation" is the process by which the server suggests safe trading stocks to users based on risk assessment.

[0578] "Notification" refers to the process by which the device notifies the user of the recommendation results sent from the server, and includes providing information via pop-up notifications or emails.

[0579] System Overview

[0580] The system of the present invention, in which a terminal and a server work in cooperation with each other, supports users in safely trading stocks. This system also incorporates an emotion engine that recognizes the user's emotions, and the results can be used to assess the risk of insider trading. Specific embodiments are described below.

[0581] Data collection by terminal

[0582] The device collects data such as the user's emails, chat messages, voice data, calendar events, and meeting minutes. This collection process is automated and runs periodically. For example, the device accesses the user's mailbox every hour to retrieve new emails and voice messages. This is done using the IMAP protocol. Additionally, calendar events are retrieved using the Google Calendar API. This data is temporarily stored in the device's local storage and then securely uploaded to the server using, for example, SFTP or HTTPS.

[0583] Data preprocessing by the server

[0584] When the server receives data sent from the device, it first preprocesses the data. HTML tags are removed from emails and chat messages and converted into pure text. This is done using Python's BeautifulSoup. Natural language processing techniques are then used to tokenize the text data and remove stop words. Voice data is converted into text using the Google Speech-to-Text API.

[0585] Emotion analysis by server

[0586] Based on the pre-processed text and audio data, the server performs sentiment analysis using an emotion engine. The text data is input into a sentiment analysis model (e.g., BERT) and an emotion label is output. The audio data is analyzed for tone and intonation to identify emotions. This emotion data is used in the subsequent risk assessment process.

[0587] Server Risk Assessment

[0588] After data preprocessing and sentiment analysis are completed, the server performs risk assessment using a machine learning model (e.g., XGBoost). It calculates an insider trading risk score based on the preprocessed text data, voice data, and sentiment data. This risk score indicates the risk level of each data.

[0589] Risk score aggregation and stock recommendations

[0590] The server uses statistical methods to compile the risk scores calculated for each data point and determine the overall risk level for each stock. Stocks that are deemed safe are sent from the server to the user's device as recommendation information. The device receives this recommendation and presents the user with a "list of stocks that are safe to buy and sell today" via a pop-up notification or email.

[0591] Specific examples

[0592] For example, if a user sends or receives an email stating, "At the next meeting, we will discuss Company A's new product," along with a voice message expressing their feelings of joy, this data is collected by the device. The device then uploads this email and voice data to the server. The server then preprocesses the data by removing HTML tags and converting it into text. The emotion engine then recognizes the user's emotion of joy. This data is then input into a machine learning model, which evaluates the risk of insider trading as high. In this case, Company A's stock, which has a high risk level, is not recommended, and instead other low-risk stocks are recommended to the user.

[0593] Prompt Sentence Examples

[0594] "Please explain in detail how the system preprocesses data collected by the device, including emails, chat messages, voice data, calendar events, and meeting minutes, and then uses an emotion engine to perform risk assessments. This ensures secure transactions."

[0595] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0596] Step 1: Data collection

[0597] The device periodically collects data such as the user's emails, chat messages, voice data, calendar events, and meeting minutes. For example, the device accesses the mail server every hour using the IMAP protocol to download new emails. It also retrieves calendar events using the Google Calendar API and saves them in local storage.

[0598] Input: Mail server, chat app, voice data source, calendar system

[0599] Output: New emails, chat messages, audio files, and calendar events stored in local storage.

[0600] Step 2: Upload data

[0601] The device uploads the collected data to a server at regular intervals, and to ensure security, the data is transferred using SFTP or HTTPS protocols.

[0602] Input: Data in local storage (emails, chat messages, audio files, calendar events)

[0603] Output: Report that data upload to server is complete

[0604] Step 3: Data Preprocessing

[0605] The server preprocesses the data received from the device. First, it uses Python's BeautifulSoup to remove HTML tags from emails and chat messages and convert them into pure text. Next, it tokenizes the text data using natural language processing techniques and removes stop words. For audio data, it converts it to text using the Google Speech-to-Text API.

[0606] Input: Uploaded emails, chat messages, audio files, calendar events

[0607] Output: Preprocessed clean text data and converted audio data

[0608] Step 4: Sentiment Analysis

[0609] The server runs an emotion engine using the preprocessed text and audio data. The emotion engine uses an emotion analysis model such as the BERT model to recognize user emotions (e.g., joy, anger, sadness) from the text data. It also analyzes tone and intonation from the audio data to recognize similar emotions.

[0610] Input: Preprocessed text data, audio data

[0611] Output: Emotional information tagged to text and audio data

[0612] Step 5: Risk assessment

[0613] The server calculates an insider trading risk score using a machine learning model (e.g., XGBoost) based on the pre-processed and sentiment-analyzed data. It integrates the pre-processed text data, voice data, and sentiment data to evaluate the risk level of each.

[0614] Input: Sentiment-analyzed text data, audio data

[0615] Output: Risk score for each data point

[0616] Step 6: Aggregate risk scores

[0617] The server then uses statistical methods to aggregate the calculated risk scores to determine the overall risk level for each security. This process involves evaluating the overall risk scores for each data point and identifying the security deemed to be the least risky.

[0618] Input: Individual Risk Score

[0619] Output: Overall risk level for each stock

[0620] Step 7: Stock Recommendation and Notification

[0621] The server recommends safe stocks to the user based on the aggregated risk level. The list of selected stocks is sent to the terminal. The terminal notifies the user via a pop-up notification or email of the "list of stocks that are safe to buy and sell today."

[0622] Input: Aggregated risk level

[0623] Output: Recommended stock notification to user device

[0624] In this way, the system of the present invention helps users to trade stocks safely.

[0625] (Application example 2)

[0626] 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."

[0627] In stock trading, there is a risk of insider trading and inappropriate trading due to users making decisions based on their emotions. Current systems perform risk assessment without taking users' emotions into account, which limits their accuracy. In addition, they lack protection against users making large transactions while in an emotional state. Therefore, there is a need for a system that performs risk assessment taking users' emotions into account and supports safe and appropriate trading.

[0628] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for converting voice data into text, means for recognizing emotions from the text data and voice data and performing risk assessment based on the emotions, and means for suspending a transaction and requesting confirmation in a high-risk situation. This enables highly accurate risk assessment that takes the user's emotions into consideration, and supports safe and appropriate transactions.

[0629] A "terminal" is a device used by a user that collects data and uploads it to a server.

[0630] The "server" is a central processing unit that receives data sent from the terminals and performs pre-processing and risk assessment.

[0631] "Communication tools" are communication means that users use on a daily basis, such as e-mail, chat applications, and voice message applications.

[0632] "Data preprocessing" refers to data processing in which the server cleans the data it receives, tokenizes it, removes stop words, and so on.

[0633] A "machine learning model" is an algorithm that is trained based on past data and used to assess risk.

[0634] "Insider trading" is the illegal act of trading stocks using inside information.

[0635] A "risk score" is a numerical representation of the risk level of a transaction, calculated using a machine learning model.

[0636] "Emotion recognition" means analyzing emotions (joy, anger, sadness, etc.) from a user's text data or voice data.

[0637] A "high-risk situation" is one in which the risk score is high and caution or restrictions on transactions are required.

[0638] "Trading products" are items such as financial products and stocks that are the subject of trading by users.

[0639] The system of the present invention evaluates insider trading risks while taking into account user sentiment and supports safe trading. Specifically, the system is composed of a terminal, a server, and a machine learning model.

[0640] Data collection and upload

[0641] The device automatically collects data from the communication tools and materials used by the user. The collected data includes emails, chat messages, voice data, calendar events, and meeting minutes. This process is performed periodically; for example, the device accesses the user's mailbox every hour to retrieve new emails. It also collects voice recordings of voice data. This data is temporarily stored in the device's local storage and periodically uploaded to the server.

[0642] Data Preprocessing

[0643] When the server receives the data sent from the device, it first preprocesses the data. This includes cleaning and standardizing the text data, tokenizing it, and removing stop words. In the case of audio data, it is converted into text using speech recognition technology. For this purpose, it is effective to use the Python SpeechRecognition library.

[0644] Emotion analysis

[0645] The server performs sentiment analysis using an emotion engine based on the preprocessed text and audio data. The emotion engine recognizes the user's emotions (happiness, anger, sadness, etc.) from the text and audio data. For text data, it uses the TextBlob library, and for audio data, it analyzes the tone and intonation of the voice.

[0646] Risk Assessment

[0647] Once preprocessing and sentiment analysis are complete, the server uses the data to perform risk assessment using a machine learning model. This model is trained on historical insider trading data and sentiment data. The machine learning model uses the scikit-learn library and applies the logistic regression algorithm.

[0648] Risk score aggregation and transaction restrictions

[0649] The risk scores calculated for each data item are aggregated by the server to determine the final risk level. If the risk level is high, the server sends a notification to the terminal requesting a pause in the transaction and requesting the user to reconfirm. This prevents users from making high-risk transactions based on emotional judgment.

[0650] Stock recommendations and notifications

[0651] The server recommends safe trading products to users based on the final risk level. This includes the process of selecting and listing low-risk trading products. The selected products are sent to the user's device and displayed via pop-up notification or email. This allows users to proceed with trading with peace of mind.

[0652] Specific examples

[0653] For example, if a user sends a voice message saying, "I want to make a big purchase right now!", the device collects this and sends it to the server, which converts the voice data into text and performs sentiment analysis. If the sentiment score is high and the transaction is deemed high risk, the server will pause the transaction and send a notification to the user requesting confirmation.

[0654] Prompt Sentence Examples

[0655] "Analyze voice messages from users expressing their desire to make a large purchase, assess the risk of electronic payment services based on that sentiment, and pause the transaction if necessary."

[0656] This allows users to avoid emotional decisions and conduct safe and appropriate transactions.

[0657] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0658] Step 1:

[0659] The device collects data from the communication tools and materials used by the user. The collection process is automated and runs periodically. Specifically, the device accesses the user's mailbox and chat application every hour to retrieve new emails and messages. The device also records the user's voice messages, which are included in the collected data. The input is the communication tool and voice data, and the output is the collected data.

[0660] Step 2:

[0661] The device uploads the collected data to the server. This is a periodic process in which data temporarily stored in the device's local storage is transferred to the server. The input is the collected data, and the output is the data transferred to the server. Specifically, the data is sent to the server using an HTTP request.

[0662] Step 3:

[0663] The server preprocesses the received data. Preprocessing includes cleaning and standardizing the text data, tokenizing it, and removing stop words. In the case of audio data, it is first converted into text using speech recognition technology. The input is the raw data uploaded to the server, and the output is the cleaned text data and the audio data converted into text. Specific operations use the SpeechRecognition and TextBlob libraries.

[0664] Step 4:

[0665] The server performs sentiment analysis based on the preprocessed text and audio data. The sentiment engine recognizes the user's sentiment from the text and audio data. The input is the preprocessed data, and the output is an emotion score. Specifically, it uses the TextBlob library to calculate the sentiment score for the text and analyzes the tone and intonation of the audio data.

[0666] Step 5:

[0667] The server performs risk assessment using a machine learning model based on the sentiment data and preprocessed data. This model is trained based on past insider trading data and sentiment data. The input is sentiment data and preprocessed data, and the output is a risk score. Specifically, it uses the scikit-learn library and applies the Logistic Regression algorithm.

[0668] Step 6:

[0669] The server aggregates the risk scores calculated for each data item and determines the final risk level. The input is multiple risk scores, and the output is the final risk level. Specifically, the scores are aggregated using a statistical aggregation method.

[0670] Step 7:

[0671] The server recommends safe trading products to the user based on the risk level. If the risk level is high, the server suspends the transaction and sends a notification to the terminal requesting confirmation. The input is the final risk level, and the output is a list of recommended trading products and a notification. Specific operations include notifying the user via a pop-up notification or email.

[0672] Step 8:

[0673] The terminal notifies the user of the recommendation results sent from the server. The input is the recommendation results and notification request from the server, and the output is a notification message to the user. Specifically, the terminal displays a "list of trading products that are safe to buy and sell today" via a pop-up notification or email.

[0674] This series of steps allows users to avoid the risks of emotional decisions and conduct safe and appropriate transactions.

[0675] 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.

[0676] 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.

[0677] 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.

[0678] [Third embodiment]

[0679] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0680] 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.

[0681] 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).

[0682] 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.

[0683] 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.

[0684] 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).

[0685] 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.

[0686] 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.

[0687] 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.

[0688] 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.

[0689] 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.

[0690] 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."

[0691] The system of the present invention, in which terminals and a server work in cooperation with each other, supports users in safely trading stocks. This system operates in the following manner: the user's terminal collects data from communication tools and documents, uploads it to the server, the server preprocesses the data, performs risk assessment, and finally recommends safe stocks to the user.

[0692] Data collection and upload

[0693] The device periodically collects data such as the user's emails, chat messages, calendar events, meeting minutes, etc. This collected data is temporarily stored on the device and then uploaded to the server at a specified time or trigger.

[0694] Data Preprocessing

[0695] When the server receives data sent from the terminal, it first preprocesses the data. Preprocessing includes standardizing the text data, tokenizing it using natural language processing, removing stop words, etc. For example, HTML tags are removed from the email body and the email is formatted as pure text.

[0696] Risk Assessment

[0697] Based on the pre-processed text data, the server performs risk assessment using a pre-trained machine learning model, which is trained on historical insider trading data and calculates an insider trading risk score for new data.

[0698] Risk score aggregation

[0699] The risk scores calculated for each data point are aggregated by the server to determine a final risk level, which is then associated with each individual stock and provides an assessment of the stock the user plans to trade.

[0700] Stock recommendations and notifications

[0701] The server recommends safe stocks to the user based on the aggregated risk level. The results are sent to the terminal and notified to the user. For example, a list of specific stocks is displayed in the form of "These are the stocks that you can safely buy and sell today."

[0702] Specific examples

[0703] For example, if a user sends or receives an email stating, "At the next board meeting, we will discuss a new product from Company X," this email is collected by the device and uploaded to the server. The server preprocesses this email and inputs it into a machine learning model to determine that it is likely to pose a high insider trading risk. In this case, the risk score will be high, and the related Company X stock will be excluded as a non-recommended stock. Meanwhile, other low-risk stocks will be recommended to the user.

[0704] This system allows users to trade stocks safely without being aware of the risk of insider trading, and also reduces legal risks for the company as a whole. Each component of this system is built by combining existing technologies, and is widely available as a feasible means.

[0705] The processing flow will be explained below.

[0706] Step 1:

[0707] The device collects data such as the user's email, calendar, meeting minutes, and chat messages. This collection process is automated and runs periodically. Specifically, the device accesses the user's mailbox at specified intervals to retrieve new emails. Calendar and meeting minutes data is also collected in the same way.

[0708] Step 2:

[0709] The device uploads the collected data to the server. The collected data is temporarily stored in local storage and then transferred to the server according to the upload schedule. At this time, the data is encrypted for security reasons.

[0710] Step 3:

[0711] The server pre-processes the data it receives. This includes cleaning the text data. Specifically, HTML tags are removed from emails and chat messages, converting them to pure text, and then tokenizing them using natural language processing to remove meaningless words (stop words).

[0712] Step 4:

[0713] The server inputs the pre-processed text data into a machine learning model, a pre-trained risk assessment model, which outputs a score indicating the degree of risk associated with insider trading for each piece of data.

[0714] Step 5:

[0715] The server aggregates multiple risk scores to determine the final risk level. Specifically, it aggregates the risk scores obtained for each data point using statistical methods to calculate an overall risk level for each stock. This allows it to evaluate whether a particular stock can be traded.

[0716] Step 6:

[0717] The server recommends safe trading stocks to users based on their risk level, which includes a process of selecting and listing stocks with low risk scores. The selected stocks are then recommended to users.

[0718] Step 7:

[0719] The terminal notifies the user of the recommendation results sent from the server. Specifically, the terminal displays a "list of stocks that are safe to buy and sell today" via a pop-up notification or email, providing the user with information to trade safely. This notification allows the user to proceed with trading with peace of mind.

[0720] Example 1

[0721] 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."

[0722] There is a need to provide an effective support system to prevent users from unintentionally causing legal problems in stock trading, which involves risks such as market manipulation and insider trading. This system must appropriately collect data from the communication methods and information used by users, perform risk assessments based on that data, and recommend safe trading assets.

[0723] 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.

[0724] In this invention, the server includes a means for preprocessing the received data, a means for evaluating a risk score associated with market manipulation using a machine learning model based on the preprocessed data, and a means for aggregating multiple risk scores to determine a final risk level, thereby enabling the recommendation of safe trading assets for users to trade stocks with peace of mind.

[0725] A "terminal" is an electronic device used by a user, which has a means of communication and the ability to collect information and upload it to a server.

[0726] "Data collection" is the process by which the terminal obtains the necessary data from the user's communication means and information.

[0727] A "server" is a central computer system that processes data received from terminals and provides functions such as risk assessment and asset recommendations.

[0728] "Data preprocessing" is the process by which the server converts the raw data it receives into an analyzable format, and specifically includes text standardization, tokenization, and removal of commonly used words.

[0729] A "machine learning model" refers to an algorithm or mathematical model that learns patterns from data and makes predictions or classifications for new data.

[0730] A "risk score" is a numerical assessment of the risk associated with market manipulation and insider trading, assessed using a machine learning model.

[0731] The "risk level" is a comprehensive index that aggregates multiple risk scores and indicates the overall degree of risk.

[0732] "Safe trading assets" refer to assets that are deemed to have a low risk level and that users can trade with confidence.

[0733] "Recommendation" refers to the act of the server recommending a transaction to the user based on the results of the risk assessment.

[0734] "Notification" is the process by which the terminal notifies the user of the recommendation results received from the server.

[0735] MODE FOR CARRYING OUT THE INVENTION

[0736] The system of the present invention aims to help users safely trade stocks, and involves the cooperation of a terminal and a server. Specifically, the terminal collects data from the user's various communication methods and information, and uploads the data to the server. The server then preprocesses the received data, performs risk assessment using a machine learning model, and recommends safe trading assets to the user. Each component and its operation are described in detail below.

[0737] Terminal data collection and upload

[0738] The device periodically scans data such as the user's emails, chat messages, calendar events, and meeting minutes. For example, a Python script installed on the device periodically checks the mailbox to detect new emails. The device accesses the user's mailbox using the Microsoft Graph API to retrieve new emails. This collected data is stored as a temporary file on the device's local disk. Then, at a specified time or triggered by a specific event, the data is uploaded to a server using the HTTPS protocol.

[0739] Data preprocessing by the server

[0740] When the server receives data sent from a terminal, it first preprocesses the data. A Flask-based web server receives the POST request and saves the data in the appropriate folder. Preprocessing includes standardizing the text data (standardizing character codes), tokenizing it using natural language processing (dividing sentences into words), and removing commonly used words (removing frequently used words such as "no" and "ni"). For example, HTML tags can be removed from the contents of meeting minutes to format them as pure text.

[0741] Risk Assessment

[0742] Based on the preprocessed data, the server uses a machine learning model to perform risk assessment. A TensorFlow model is used to score the likelihood of insider trading. This score is evaluated on a scale of 0 to 100; for example, an email containing a discussion about a new product from X Corporation would be calculated as a risk score of 90. The machine learning model is trained based on past market manipulation data, allowing it to quickly and accurately assess risk on new data.

[0743] Risk score aggregation

[0744] The risk scores obtained for each data item are aggregated by the server. Higher scores are excluded from trading, and lower scores are listed. For example, if "Company X's risk score is 90" and "Company Y's risk score is 20," then "Company Y" will be included in the recommendation list.

[0745] Stock recommendations and notifications

[0746] Based on the aggregated risk scores, the server recommends safe trading assets to the user. The recommendation results are sent to the terminal, which notifies the user through a notification system. For example, a specific list of stocks may be displayed, such as "The following stocks are safe for you to buy and sell today: Y Co., Ltd." The terminal uses the notification function to provide information to the user in real time.

[0747] Examples of concrete examples and prompts

[0748] For example, if a user sends or receives an email with the content "We will discuss a new product at the next board meeting," this email is collected by the device and uploaded to the server. The server preprocesses this email and inputs it into a machine learning model, which may determine that the email poses a high insider trading risk. In this case, the risk score will be high and the related stocks will not be recommended. Instead, other low-risk stocks will be recommended to the user.

[0749] Below are some example prompts for generative AI models:

[0750] "We have detected an email containing information about a new product that the user would like to discuss at the next board meeting. Based on this information, we would like you to calculate an insider trading risk score and recommend safe stocks to trade."

[0751] In this way, the system of the present invention works in conjunction with various communication means to assist users in safely and efficiently trading stocks without unintentionally incurring legal risks.

[0752] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0753] Step 1:

[0754] Data collection

[0755] The device periodically collects data from the user's emails, chat messages, calendar events, meeting minutes, etc. For example, a Python script installed on the device uses the Microsoft Graph API to access the user's mailbox and retrieve new emails.

[0756] Input: Various user communication data (emails, chat messages, etc.)

[0757] Output: Raw data stored on the device

[0758] Step 2:

[0759] Data storage and trigger settings

[0760] The collected data is saved as a temporary file on the device's local disk. For example, meeting minutes are saved as "meeting_notes.txt." The device also uploads data at a specified time (e.g., midnight) or when a specific event (e.g., receiving an email) triggers the upload.

[0761] Input: Raw data stored on the device

[0762] Output: Prepared upload data

[0763] Step 3:

[0764] Data upload

[0765] The data is encrypted and uploaded to the server using the HTTPS protocol: the script on the device sends the collected data to the server using HTTP POST requests.

[0766] Input: Prepared upload data

[0767] Output: Data sent to the server

[0768] Step 4:

[0769] Data reception

[0770] The server receives the data sent from the device. For example, a Flask-based web server receives a POST request and saves the data in the appropriate folder (e.g., " / data / incoming").

[0771] Input: Data sent from the terminal

[0772] Output: Temporary data on the server

[0773] Step 5:

[0774] Data Preprocessing

[0775] The server performs preprocessing on the received data, including standardization, tokenization, and stop word removal. Specifically, it uses Python's Beautiful Soup library to remove HTML tags, and the NLTK library to tokenize each word and remove frequent words. For example, it removes HTML tags from the content of "meeting_notes.txt" and converts it into pure text.

[0776] Input: Temporary data on the server

[0777] Output: Preprocessed data

[0778] Step 6:

[0779] Risk Assessment

[0780] Based on the preprocessed data, the server uses machine learning models to perform risk assessments. For example, TensorFlow can be used to calculate a risk score that indicates the likelihood of insider trading. For example, minutes of a meeting about a new product might be assigned a high risk score (e.g., 90).

[0781] Input: Preprocessed data

[0782] Output: Risk score

[0783] Step 7:

[0784] Risk score aggregation

[0785] The server aggregates the risk scores obtained for each data. For example, if "Company A's risk score is 90" and "Company B's risk score is 20," then "Company B" will be included in the recommendation list.

[0786] Input: Risk Score

[0787] Output: Aggregated risk level

[0788] Step 8:

[0789] Stock Recommendations

[0790] The server recommends safe trading assets to users based on the aggregated risk level. Specifically, it generates a recommended list of assets with low risk levels.

[0791] Input: Aggregated risk level

[0792] Output: Recommended stock list

[0793] Step 9:

[0794] Notification of recommendation results

[0795] The server sends the recommendation results to the terminal, which then notifies the user. Specifically, the terminal's notification system is used to provide real-time information to the user. For example, it may notify the user that "This stock is safe to trade today: Co. B."

[0796] Input: Recommended stock list

[0797] Output: User notification

[0798] By following the steps above, users can safely trade stocks while minimizing the risk of insider trading.

[0799] (Application example 1)

[0800] 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."

[0801] Currently, it is extremely important to properly manage insider trading risks in stock trading. However, existing systems inadequately assess the risk of stocks traded by users, and manual confirmation work requires a great deal of effort. Furthermore, even when safe stocks are recommended, the procedures for immediate purchase are cumbersome, hindering smooth trading. Therefore, the present invention aims to provide a system that supports users in safely and efficiently trading stocks, enabling risk assessment and immediate purchase.

[0802] 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.

[0803] In this invention, the server includes means for evaluating a risk score associated with insider trading using a machine learning model based on the preprocessed data, means for aggregating multiple risk scores to determine a final risk level, and means for instantly purchasing safe stocks in cooperation with the user's electronic payment platform based on the aggregated risk level, thereby enabling the user to quickly purchase safe trading stocks.

[0804] A "terminal" is an electronic device used by a user, such as a smartphone or PC, that collects and notifies data.

[0805] A "server" is a computer system that receives, processes, and evaluates data sent from a terminal.

[0806] "Communication tools" are tools used by users, such as email, chat messages, and calendar events.

[0807] "Materials" refers to document data related to a transaction, such as a user's electronic documents or meeting minutes.

[0808] "Data collection" is the process by which the device periodically collects data from the user's communication tools and materials.

[0809] "Data preprocessing" is the process by which the server standardizes, tokenizes, removes stop words, etc., data sent from the terminal.

[0810] A "machine learning model" is a model that is trained based on past insider trading data and performs risk assessments on new data.

[0811] The "risk score" is a numerical representation of insider trading risk calculated using a machine learning model.

[0812] The "risk level" is an index that indicates the safety of a trading stock and is determined by aggregating multiple risk scores.

[0813] "Trading assets" are financial assets such as stocks and securities that users buy and sell.

[0814] "Recommendation" refers to the act of the server suggesting trading assets to a user based on risk level.

[0815] "Notification" is the process by which the terminal notifies the user of the recommendation results received from the server.

[0816] An "electronic payment platform" is an online payment system used to purchase and sell trading assets.

[0817] "Linkage" is the process by which the server communicates data with the electronic payment platform to realize the transaction.

[0818] The system of the present invention involves a terminal and a server working together to evaluate the risk of insider trading in stock trading and recommend safe stocks to users. The components of this system and their operation are described below.

[0819] Data collection

[0820] The device periodically collects data from the communication tools (emails, chat messages, calendar events, etc.) and documents used by the user. This collected data is temporarily stored on the device.

[0821] Uploading data

[0822] At a specified time or trigger, the collected data is consolidated and uploaded to the server by the device, and the uploaded data is securely transmitted to the server via the Internet.

[0823] Data Preprocessing

[0824] The server preprocesses the received data, which includes standardizing the text data (removing HTML tags, etc.), tokenizing (splitting words), removing stop words (removing non-important words), etc. Spacy is used as the natural language processing library.

[0825] Risk Assessment

[0826] Based on the preprocessed text data, the server uses a generative AI model (machine learning model) to calculate a risk score. This model is trained on past insider trading data and assesses the insider trading risk of new data.

[0827] Determining the risk level

[0828] The server aggregates multiple risk scores and determines a final risk level, which evaluates the overall insider trading risk for each stock.

[0829] Stock recommendation and electronic payment platform integration

[0830] The server recommends safe trading assets to users based on the aggregated risk level, and also provides a means for users to instantly purchase safe stocks in conjunction with their electronic payment platform, allowing users to quickly purchase the recommended stocks.

[0831] Notification of recommendation results

[0832] Finally, the server sends the recommendation results to the device, which then notifies the user via a smartphone application or a PC notification system.

[0833] Specific examples

[0834] For example, if a user sends or receives an email with the content "At the next meeting, we will discuss a new product from Company X," this email is collected by the device and uploaded to the server. The server preprocesses this email and inputs it into a machine learning model to determine that it poses a high risk of insider trading. In this case, the risk score will be high, and the related Company X's stock will be excluded as a non-recommended stock. Meanwhile, other low-risk stocks will be recommended to the user.

[0835] An example of a prompt sentence to input to the generative AI model is as follows:

[0836] At the next meeting, we will discuss Company X's new product.

[0837] Yu

[0838] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0839] Step 1:

[0840] The device collects data from the user's communication tools and documents. Input data includes emails, chat messages, calendar events, etc., and periodically scans them to extract important information and temporarily store it on the device. The output is a list of the collected data.

[0841] Step 2:

[0842] The device uploads the collected data to the server. Here, the data is encrypted and securely sent to the server at regular intervals or when a specific trigger (user-specified operation or detection of new data) is triggered. The input is the data stored on the device, and the output is the data received by the server.

[0843] Step 3:

[0844] The server preprocesses the received data. Specifically, it standardizes, tokenizes, and removes stop words. Spacy is used as a natural language processing library, and the input data is text data, and the output is preprocessed, cleaned text data.

[0845] Step 4:

[0846] The server evaluates a risk score associated with insider trading using a generative AI model based on the preprocessed data. The machine learning model is trained on historical insider trading data and takes the preprocessed text data as input. The output is a calculated insider trading risk score.

[0847] Step 5:

[0848] The server aggregates multiple risk scores and determines the final risk level. Here, the risk scores for each data are integrated to perform an overall risk assessment. The input is the individual risk scores, and the output is the integrated final risk level.

[0849] Step 6:

[0850] The server recommends safe trading assets to users based on their risk level. Depending on the risk level, stocks deemed safe are selected and a list is generated. The input is the final risk level, and the output is a list of recommended safe trading assets.

[0851] Step 7:

[0852] The server instantly purchases safe stocks based on the aggregated risk level in conjunction with the user's electronic payment platform. The server uses the user's account information and payment system to automate the purchase process for the recommended stocks. The input is a list of recommended trading assets, and the output is confirmation of the completion of the purchase process.

[0853] Step 8:

[0854] The terminal notifies the user of the recommendation results and purchase completion notification. Specifically, the recommended stocks and purchase results are notified to the user via a smartphone application or a PC notification system. The input is the notification content sent from the server, and the output is what is displayed to the user.

[0855] 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.

[0856] The system of the present invention operates in cooperation with a terminal and a server to support users in safely trading stocks. This system also incorporates an emotion engine that recognizes the user's emotions, and the results can be used to assess the risk of insider trading. Specific embodiments are described below.

[0857] Data collection and upload

[0858] The device collects data such as the user's emails, chat messages, voice data, calendar events, and meeting minutes. This collection process is automated and runs periodically. For example, the device accesses the user's mailbox every hour to retrieve new emails and voice messages. Calendar and meeting minutes data is also retrieved in the same way. This data is temporarily stored in the device's local storage and periodically uploaded to the server.

[0859] Data Preprocessing

[0860] When the server receives data sent from a terminal, it first preprocesses the data. Preprocessing includes cleaning the text data. Specifically, HTML tags are removed from emails and chat messages to convert them into pure text. Natural language processing is then used to tokenize the data and remove meaningless words (stop words). In the case of audio data, speech recognition technology is used to convert it into text.

[0861] Emotion analysis

[0862] The server uses the preprocessed text and voice data to perform emotion analysis using an emotion engine. The emotion engine recognizes the user's emotions (happiness, anger, sadness, etc.) from the text data. It also analyzes the tone and intonation of the voice data to recognize similar emotions. This emotion data is then incorporated into the risk assessment process.

[0863] Risk Assessment

[0864] Once preprocessing and sentiment analysis are complete, the server uses a machine learning model to perform risk assessment on the data. The model analyzes the preprocessed text data, audio data, and sentiment data together to calculate a risk score related to insider trading. This score indicates the risk level of each data.

[0865] Risk score aggregation

[0866] The risk scores calculated for each data item are aggregated by the server, and the final risk level is determined. Specifically, the risk scores obtained for each data item are aggregated using statistical methods to calculate the overall risk level for each stock. This allows an evaluation of whether a particular stock can be traded.

[0867] Stock recommendations and notifications

[0868] The server recommends safe stocks to the user based on the aggregated risk level. This involves the process of selecting and listing stocks with low risk scores. The selected stocks are recommended to the user and ultimately sent to the user's device. The user's device receives the recommendation results and notifies the user. The device displays a pop-up notification or email with a "list of stocks that are safe to buy and sell today," providing the user with information to trade safely. This notification allows the user to proceed with trading with peace of mind.

[0869] Specific examples

[0870] For example, if a user sends and receives an email stating, "At the next board meeting, we will discuss a new product from Company X," along with a voice message expressing joy, this data is collected by the device and uploaded to the server. The server preprocesses this email and voice data and uses an emotion engine to recognize that the user is expressing joy. The server then inputs this data into a machine learning model and determines that the insider trading risk is high (high risk score). In this case, shares of Company X, which has a high risk level, are not recommended to the user, and instead other low-risk stocks are recommended to the user.

[0871] This system allows users to trade stocks safely without worrying about the risk of insider trading, and reduces legal risks for the company as a whole. The addition of an emotion engine enables risk assessment that takes user emotions into account, resulting in more accurate recommendations. Each component of this system is built by combining existing technologies, making it a widely applicable and feasible solution.

[0872] The processing flow will be explained below.

[0873] Step 1:

[0874] The device collects data such as the user's emails, chat messages, voice data, calendar events, and meeting minutes. This collection process is automated, so for example, the device accesses the user's mailbox every 30 minutes to retrieve new emails and newly recorded voice data. Calendar information and meeting minutes data are also collected in the same way.

[0875] Step 2:

[0876] The device uploads the collected data to the server. The collected data is temporarily stored in the device's local storage and then transferred to the server based on a set schedule. At this time, the data is encrypted to ensure security.

[0877] Step 3:

[0878] The server preprocesses the data it receives. Preprocessing includes cleaning the text data. For example, HTML tags are removed from emails and chat messages to convert them into pure text. Natural language processing technology is then used to tokenize the data and remove meaningless words (stop words). In the case of audio data, speech recognition technology is used to convert it into text.

[0879] Step 4:

[0880] The server uses an emotion engine to analyze the user's emotions from the preprocessed text and voice data. The emotion engine uses, for example, natural language processing (NLP) techniques to recognize the user's emotions (happiness, anger, sadness, etc.) from the text data. It also analyzes the tone and intonation of the voice data to recognize similar emotions. The recognized emotion data is then incorporated into the subsequent risk assessment process.

[0881] Step 5:

[0882] The server uses the sentiment-analyzed data to perform risk assessment using a machine learning model. The machine learning model is a trained model that calculates a risk score related to insider trading. The model takes the preprocessed text data, audio data, and sentiment data as input and outputs a numerical risk level for each data.

[0883] Step 6:

[0884] The server aggregates multiple risk scores to determine the final risk level. Specifically, the risk scores obtained for each data point are aggregated using statistical methods such as weighted averaging to calculate an overall risk level for each stock. This is used to evaluate whether a particular stock can be traded.

[0885] Step 7:

[0886] The server recommends safe trading stocks to the user based on the risk level. Stocks with low risk scores are selected and listed. This recommendation list is sent from the server to the user's terminal.

[0887] Step 8:

[0888] The terminal notifies the user of the recommendation results sent from the server. Specifically, the terminal displays a "list of stocks that are safe to buy and sell today" via a pop-up notification or email, providing the user with information to trade safely. This notification allows the user to trade with peace of mind.

[0889] Example 2

[0890] 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."

[0891] The challenge is to provide a system that reduces the risk of insider trading and supports safe stock trading. Conventional systems do not take user sentiment or unstructured data into account, which can result in low accuracy in risk assessment and the possibility of incorrect trade recommendations. Furthermore, data preprocessing and risk assessment require a significant amount of time and effort, so automation and improved accuracy are required.

[0892] 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.

[0893] In this invention, the server includes means for preprocessing received data, means for analyzing user emotions based on the preprocessed data using an emotion recognition engine, and means for evaluating a risk score related to insider trading using a machine learning model based on the emotion-analyzed data. This enables highly accurate risk assessment that takes user emotions and unstructured data into consideration, and the recommendation of safe stocks to trade.

[0894] "Terminal" means a computing device used by a user, primarily a hardware device or software application for data collection, processing, and notification.

[0895] "Server" refers to a computer system that receives data sent from terminals and performs centralized processing such as pre-processing, sentiment analysis, risk assessment, and stock recommendations.

[0896] An "emotion recognition engine" is a machine learning model or algorithm that analyzes a user's emotions from text and voice data and identifies the type and intensity of those emotions.

[0897] A "machine learning model" is an algorithm that learns patterns and trends based on past data and makes predictions and classifications for new data, and is used particularly for risk assessment in the present invention.

[0898] A "risk score" is a numerical representation of the degree of risk associated with insider trading, and is an index calculated for each data point.

[0899] "Preprocessing" is the process of applying initial processing to received data, including text standardization, tokenization, stop word removal, and converting audio data to text.

[0900] "Communication Tools" means software or platforms that allow users to send and receive messages and make voice and video calls, such as email clients and chat applications.

[0901] An "information source" is a source of digital content that a user accesses, such as a mail server, chat application, or calendar system.

[0902] "Uploading" is the process by which a device sends collected data to a server, using a secure transfer protocol such as SFTP or HTTPS.

[0903] "Stock recommendation" is the process by which the server suggests safe trading stocks to users based on risk assessment.

[0904] "Notification" refers to the process by which the device notifies the user of the recommendation results sent from the server, and includes providing information via pop-up notifications or emails.

[0905] System Overview

[0906] The system of the present invention, in which a terminal and a server work in cooperation with each other, supports users in safely trading stocks. This system also incorporates an emotion engine that recognizes the user's emotions, and the results can be used to assess the risk of insider trading. Specific embodiments are described below.

[0907] Data collection by terminal

[0908] The device collects data such as the user's emails, chat messages, voice data, calendar events, and meeting minutes. This collection process is automated and runs periodically. For example, the device accesses the user's mailbox every hour to retrieve new emails and voice messages. This is done using the IMAP protocol. Additionally, calendar events are retrieved using the Google Calendar API. This data is temporarily stored in the device's local storage and then securely uploaded to the server using, for example, SFTP or HTTPS.

[0909] Data preprocessing by the server

[0910] When the server receives data sent from the device, it first preprocesses the data. HTML tags are removed from emails and chat messages and converted into pure text. This is done using Python's BeautifulSoup. Natural language processing techniques are then used to tokenize the text data and remove stop words. Voice data is converted into text using the Google Speech-to-Text API.

[0911] Emotion analysis by server

[0912] Based on the pre-processed text and audio data, the server performs sentiment analysis using an emotion engine. The text data is input into a sentiment analysis model (e.g., BERT) and an emotion label is output. The audio data is analyzed for tone and intonation to identify emotions. This emotion data is used in the subsequent risk assessment process.

[0913] Server Risk Assessment

[0914] After data preprocessing and sentiment analysis are completed, the server performs risk assessment using a machine learning model (e.g., XGBoost). It calculates an insider trading risk score based on the preprocessed text data, voice data, and sentiment data. This risk score indicates the risk level of each data.

[0915] Risk score aggregation and stock recommendations

[0916] The server uses statistical methods to compile the risk scores calculated for each data point and determine the overall risk level for each stock. Stocks that are deemed safe are sent from the server to the user's device as recommendation information. The device receives this recommendation and presents the user with a "list of stocks that are safe to buy and sell today" via a pop-up notification or email.

[0917] Specific examples

[0918] For example, if a user sends or receives an email stating, "At the next meeting, we will discuss Company A's new product," along with a voice message expressing their feelings of joy, this data is collected by the device. The device then uploads this email and voice data to the server. The server then preprocesses the data by removing HTML tags and converting it into text. The emotion engine then recognizes the user's emotion of joy. This data is then input into a machine learning model, which evaluates the risk of insider trading as high. In this case, Company A's stock, which has a high risk level, is not recommended, and instead other low-risk stocks are recommended to the user.

[0919] Prompt Sentence Examples

[0920] "Please explain in detail how the system preprocesses data collected by the device, including emails, chat messages, voice data, calendar events, and meeting minutes, and then uses an emotion engine to perform risk assessments. This ensures secure transactions."

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

[0922] Step 1: Data collection

[0923] The device periodically collects data such as the user's emails, chat messages, voice data, calendar events, and meeting minutes. For example, the device accesses the mail server every hour using the IMAP protocol to download new emails. It also retrieves calendar events using the Google Calendar API and saves them in local storage.

[0924] Input: Mail server, chat app, voice data source, calendar system

[0925] Output: New emails, chat messages, audio files, and calendar events stored in local storage.

[0926] Step 2: Upload data

[0927] The device uploads the collected data to a server at regular intervals, and to ensure security, the data is transferred using SFTP or HTTPS protocols.

[0928] Input: Data in local storage (emails, chat messages, audio files, calendar events)

[0929] Output: Report that data upload to server is complete

[0930] Step 3: Data Preprocessing

[0931] The server preprocesses the data received from the device. First, it uses Python's BeautifulSoup to remove HTML tags from emails and chat messages and convert them into pure text. Next, it tokenizes the text data using natural language processing techniques and removes stop words. For audio data, it converts it to text using the Google Speech-to-Text API.

[0932] Input: Uploaded emails, chat messages, audio files, calendar events

[0933] Output: Preprocessed clean text data and converted audio data

[0934] Step 4: Sentiment Analysis

[0935] The server runs an emotion engine using the preprocessed text and audio data. The emotion engine uses an emotion analysis model such as the BERT model to recognize user emotions (e.g., joy, anger, sadness) from the text data. It also analyzes tone and intonation from the audio data to recognize similar emotions.

[0936] Input: Preprocessed text data, audio data

[0937] Output: Emotional information tagged to text and audio data

[0938] Step 5: Risk assessment

[0939] The server calculates an insider trading risk score using a machine learning model (e.g., XGBoost) based on the pre-processed and sentiment-analyzed data. It integrates the pre-processed text data, voice data, and sentiment data to evaluate the risk level of each.

[0940] Input: Sentiment-analyzed text data, audio data

[0941] Output: Risk score for each data point

[0942] Step 6: Aggregate risk scores

[0943] The server then uses statistical methods to aggregate the calculated risk scores to determine the overall risk level for each security. This process involves evaluating the overall risk scores for each data point and identifying the security deemed to be the least risky.

[0944] Input: Individual Risk Score

[0945] Output: Overall risk level for each stock

[0946] Step 7: Stock Recommendation and Notification

[0947] The server recommends safe stocks to the user based on the aggregated risk level. The list of selected stocks is sent to the terminal. The terminal notifies the user via a pop-up notification or email of the "list of stocks that are safe to buy and sell today."

[0948] Input: Aggregated risk level

[0949] Output: Recommended stock notification to user device

[0950] In this way, the system of the present invention helps users to trade stocks safely.

[0951] (Application example 2)

[0952] 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."

[0953] In stock trading, there is a risk of insider trading and inappropriate trading due to users making decisions based on their emotions. Current systems perform risk assessment without taking users' emotions into account, which limits their accuracy. In addition, they lack protection against users making large transactions while in an emotional state. Therefore, there is a need for a system that performs risk assessment taking users' emotions into account and supports safe and appropriate trading.

[0954] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for converting voice data into text, means for recognizing emotions from the text data and voice data and performing risk assessment based on the emotions, and means for suspending a transaction and requesting confirmation in a high-risk situation. This enables highly accurate risk assessment that takes the user's emotions into consideration, and supports safe and appropriate transactions.

[0955] A "terminal" is a device used by a user that collects data and uploads it to a server.

[0956] The "server" is a central processing unit that receives data sent from the terminals and performs pre-processing and risk assessment.

[0957] "Communication tools" are communication means that users use on a daily basis, such as e-mail, chat applications, and voice message applications.

[0958] "Data preprocessing" refers to data processing in which the server cleans the data it receives, tokenizes it, removes stop words, and so on.

[0959] A "machine learning model" is an algorithm that is trained based on past data and used to assess risk.

[0960] "Insider trading" is the illegal act of trading stocks using inside information.

[0961] A "risk score" is a numerical representation of the risk level of a transaction, calculated using a machine learning model.

[0962] "Emotion recognition" means analyzing emotions (joy, anger, sadness, etc.) from a user's text data or voice data.

[0963] A "high-risk situation" is one in which the risk score is high and caution or restrictions on transactions are required.

[0964] "Trading products" are items such as financial products and stocks that are the subject of trading by users.

[0965] The system of the present invention evaluates insider trading risks while taking into account user sentiment and supports safe trading. Specifically, the system is composed of a terminal, a server, and a machine learning model.

[0966] Data collection and upload

[0967] The device automatically collects data from the communication tools and materials used by the user. The collected data includes emails, chat messages, voice data, calendar events, and meeting minutes. This process is performed periodically; for example, the device accesses the user's mailbox every hour to retrieve new emails. It also collects voice recordings of voice data. This data is temporarily stored in the device's local storage and periodically uploaded to the server.

[0968] Data Preprocessing

[0969] When the server receives the data sent from the device, it first preprocesses the data. This includes cleaning and standardizing the text data, tokenizing it, and removing stop words. In the case of audio data, it is converted into text using speech recognition technology. For this purpose, it is effective to use the Python SpeechRecognition library.

[0970] Emotion analysis

[0971] The server performs sentiment analysis using an emotion engine based on the preprocessed text and audio data. The emotion engine recognizes the user's emotions (happiness, anger, sadness, etc.) from the text and audio data. For text data, it uses the TextBlob library, and for audio data, it analyzes the tone and intonation of the voice.

[0972] Risk Assessment

[0973] Once preprocessing and sentiment analysis are complete, the server uses the data to perform risk assessment using a machine learning model. This model is trained on historical insider trading data and sentiment data. The machine learning model uses the scikit-learn library and applies the logistic regression algorithm.

[0974] Risk score aggregation and transaction restrictions

[0975] The risk scores calculated for each data item are aggregated by the server to determine the final risk level. If the risk level is high, the server sends a notification to the terminal requesting a pause in the transaction and requesting the user to reconfirm. This prevents users from making high-risk transactions based on emotional judgment.

[0976] Stock recommendations and notifications

[0977] The server recommends safe trading products to users based on the final risk level. This includes the process of selecting and listing low-risk trading products. The selected products are sent to the user's device and displayed via pop-up notification or email. This allows users to proceed with trading with peace of mind.

[0978] Specific examples

[0979] For example, if a user sends a voice message saying, "I want to make a big purchase right now!", the device collects this and sends it to the server, which converts the voice data into text and performs sentiment analysis. If the sentiment score is high and the transaction is deemed high risk, the server will pause the transaction and send a notification to the user requesting confirmation.

[0980] Prompt Sentence Examples

[0981] "Analyze voice messages from users expressing their desire to make a large purchase, assess the risk of electronic payment services based on that sentiment, and pause the transaction if necessary."

[0982] This allows users to avoid emotional decisions and conduct safe and appropriate transactions.

[0983] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0984] Step 1:

[0985] The device collects data from the communication tools and materials used by the user. The collection process is automated and runs periodically. Specifically, the device accesses the user's mailbox and chat application every hour to retrieve new emails and messages. The device also records the user's voice messages, which are included in the collected data. The input is the communication tool and voice data, and the output is the collected data.

[0986] Step 2:

[0987] The device uploads the collected data to the server. This is a periodic process in which data temporarily stored in the device's local storage is transferred to the server. The input is the collected data, and the output is the data transferred to the server. Specifically, the data is sent to the server using an HTTP request.

[0988] Step 3:

[0989] The server preprocesses the received data. Preprocessing includes cleaning and standardizing the text data, tokenizing it, and removing stop words. In the case of audio data, it is first converted into text using speech recognition technology. The input is the raw data uploaded to the server, and the output is the cleaned text data and the audio data converted into text. Specific operations use the SpeechRecognition and TextBlob libraries.

[0990] Step 4:

[0991] The server performs sentiment analysis based on the preprocessed text and audio data. The sentiment engine recognizes the user's sentiment from the text and audio data. The input is the preprocessed data, and the output is an emotion score. Specifically, it uses the TextBlob library to calculate the sentiment score for the text and analyzes the tone and intonation of the audio data.

[0992] Step 5:

[0993] The server performs risk assessment using a machine learning model based on the sentiment data and preprocessed data. This model is trained based on past insider trading data and sentiment data. The input is sentiment data and preprocessed data, and the output is a risk score. Specifically, it uses the scikit-learn library and applies the Logistic Regression algorithm.

[0994] Step 6:

[0995] The server aggregates the risk scores calculated for each data item and determines the final risk level. The input is multiple risk scores, and the output is the final risk level. Specifically, the scores are aggregated using a statistical aggregation method.

[0996] Step 7:

[0997] The server recommends safe trading products to the user based on the risk level. If the risk level is high, the server suspends the transaction and sends a notification to the terminal requesting confirmation. The input is the final risk level, and the output is a list of recommended trading products and a notification. Specific operations include notifying the user via a pop-up notification or email.

[0998] Step 8:

[0999] The terminal notifies the user of the recommendation results sent from the server. The input is the recommendation results and notification request from the server, and the output is a notification message to the user. Specifically, the terminal displays a "list of trading products that are safe to buy and sell today" via a pop-up notification or email.

[1000] This series of steps allows users to avoid the risks of emotional decisions and conduct safe and appropriate transactions.

[1001] 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.

[1002] 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.

[1003] 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.

[1004] [Fourth embodiment]

[1005] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1006] 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.

[1007] 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).

[1008] 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.

[1009] 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.

[1010] 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).

[1011] 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.

[1012] 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.

[1013] 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.

[1014] 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.

[1015] 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.

[1016] 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.

[1017] 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."

[1018] The system of the present invention, in which terminals and a server work in cooperation with each other, supports users in safely trading stocks. This system operates in the following manner: the user's terminal collects data from communication tools and documents, uploads it to the server, the server preprocesses the data, performs risk assessment, and finally recommends safe stocks to the user.

[1019] Data collection and upload

[1020] The device periodically collects data such as the user's emails, chat messages, calendar events, meeting minutes, etc. This collected data is temporarily stored on the device and then uploaded to the server at a specified time or trigger.

[1021] Data Preprocessing

[1022] When the server receives data sent from the terminal, it first preprocesses the data. Preprocessing includes standardizing the text data, tokenizing it using natural language processing, removing stop words, etc. For example, HTML tags are removed from the email body and the email is formatted as pure text.

[1023] Risk Assessment

[1024] Based on the pre-processed text data, the server performs risk assessment using a pre-trained machine learning model, which is trained on historical insider trading data and calculates an insider trading risk score for new data.

[1025] Risk score aggregation

[1026] The risk scores calculated for each data point are aggregated by the server to determine a final risk level, which is then associated with each individual stock and provides an assessment of the stock the user plans to trade.

[1027] Stock recommendations and notifications

[1028] The server recommends safe stocks to the user based on the aggregated risk level. The results are sent to the terminal and notified to the user. For example, a list of specific stocks is displayed in the form of "These are the stocks that you can safely buy and sell today."

[1029] Specific examples

[1030] For example, if a user sends or receives an email stating, "At the next board meeting, we will discuss a new product from Company X," this email is collected by the device and uploaded to the server. The server preprocesses this email and inputs it into a machine learning model to determine that it is likely to pose a high insider trading risk. In this case, the risk score will be high, and the related Company X stock will be excluded as a non-recommended stock. Meanwhile, other low-risk stocks will be recommended to the user.

[1031] This system allows users to trade stocks safely without being aware of the risk of insider trading, and also reduces legal risks for the company as a whole. Each component of this system is built by combining existing technologies, and is widely available as a feasible means.

[1032] The processing flow will be explained below.

[1033] Step 1:

[1034] The device collects data such as the user's email, calendar, meeting minutes, and chat messages. This collection process is automated and runs periodically. Specifically, the device accesses the user's mailbox at specified intervals to retrieve new emails. Calendar and meeting minutes data is also collected in the same way.

[1035] Step 2:

[1036] The device uploads the collected data to the server. The collected data is temporarily stored in local storage and then transferred to the server according to the upload schedule. At this time, the data is encrypted for security reasons.

[1037] Step 3:

[1038] The server pre-processes the data it receives. This includes cleaning the text data. Specifically, HTML tags are removed from emails and chat messages, converting them to pure text, and then tokenizing them using natural language processing to remove meaningless words (stop words).

[1039] Step 4:

[1040] The server inputs the pre-processed text data into a machine learning model, a pre-trained risk assessment model, which outputs a score indicating the degree of risk associated with insider trading for each piece of data.

[1041] Step 5:

[1042] The server aggregates multiple risk scores to determine the final risk level. Specifically, it aggregates the risk scores obtained for each data point using statistical methods to calculate an overall risk level for each stock. This allows it to evaluate whether a particular stock can be traded.

[1043] Step 6:

[1044] The server recommends safe trading stocks to users based on their risk level, which includes a process of selecting and listing stocks with low risk scores. The selected stocks are then recommended to users.

[1045] Step 7:

[1046] The terminal notifies the user of the recommendation results sent from the server. Specifically, the terminal displays a "list of stocks that are safe to buy and sell today" via a pop-up notification or email, providing the user with information to trade safely. This notification allows the user to proceed with trading with peace of mind.

[1047] Example 1

[1048] 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."

[1049] There is a need to provide an effective support system to prevent users from unintentionally causing legal problems in stock trading, which involves risks such as market manipulation and insider trading. This system must appropriately collect data from the communication methods and information used by users, perform risk assessments based on that data, and recommend safe trading assets.

[1050] 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.

[1051] In this invention, the server includes a means for preprocessing the received data, a means for evaluating a risk score associated with market manipulation using a machine learning model based on the preprocessed data, and a means for aggregating multiple risk scores to determine a final risk level, thereby enabling the recommendation of safe trading assets for users to trade stocks with peace of mind.

[1052] A "terminal" is an electronic device used by a user, which has a means of communication and the ability to collect information and upload it to a server.

[1053] "Data collection" is the process by which the terminal obtains the necessary data from the user's communication means and information.

[1054] A "server" is a central computer system that processes data received from terminals and provides functions such as risk assessment and asset recommendations.

[1055] "Data preprocessing" is the process by which the server converts the raw data it receives into an analyzable format, and specifically includes text standardization, tokenization, and removal of commonly used words.

[1056] A "machine learning model" refers to an algorithm or mathematical model that learns patterns from data and makes predictions or classifications for new data.

[1057] A "risk score" is a numerical assessment of the risk associated with market manipulation and insider trading, assessed using a machine learning model.

[1058] The "risk level" is a comprehensive index that aggregates multiple risk scores and indicates the overall degree of risk.

[1059] "Safe trading assets" refer to assets that are deemed to have a low risk level and that users can trade with confidence.

[1060] "Recommendation" refers to the act of the server recommending a transaction to the user based on the results of the risk assessment.

[1061] "Notification" is the process by which the terminal notifies the user of the recommendation results received from the server.

[1062] MODE FOR CARRYING OUT THE INVENTION

[1063] The system of the present invention aims to help users safely trade stocks, and involves the cooperation of a terminal and a server. Specifically, the terminal collects data from the user's various communication methods and information, and uploads the data to the server. The server then preprocesses the received data, performs risk assessment using a machine learning model, and recommends safe trading assets to the user. Each component and its operation are described in detail below.

[1064] Terminal data collection and upload

[1065] The device periodically scans data such as the user's emails, chat messages, calendar events, and meeting minutes. For example, a Python script installed on the device periodically checks the mailbox to detect new emails. The device accesses the user's mailbox using the Microsoft Graph API to retrieve new emails. This collected data is stored as a temporary file on the device's local disk. Then, at a specified time or triggered by a specific event, the data is uploaded to a server using the HTTPS protocol.

[1066] Data preprocessing by the server

[1067] When the server receives data sent from a terminal, it first preprocesses the data. A Flask-based web server receives the POST request and saves the data in the appropriate folder. Preprocessing includes standardizing the text data (standardizing character codes), tokenizing it using natural language processing (dividing sentences into words), and removing commonly used words (removing frequently used words such as "no" and "ni"). For example, HTML tags can be removed from the contents of meeting minutes to format them as pure text.

[1068] Risk Assessment

[1069] Based on the preprocessed data, the server uses a machine learning model to perform risk assessment. A TensorFlow model is used to score the likelihood of insider trading. This score is evaluated on a scale of 0 to 100; for example, an email containing a discussion about a new product from X Corporation would be calculated as a risk score of 90. The machine learning model is trained based on past market manipulation data, allowing it to quickly and accurately assess risk on new data.

[1070] Risk score aggregation

[1071] The risk scores obtained for each data item are aggregated by the server. Higher scores are excluded from trading, and lower scores are listed. For example, if "Company X's risk score is 90" and "Company Y's risk score is 20," then "Company Y" will be included in the recommendation list.

[1072] Stock recommendations and notifications

[1073] Based on the aggregated risk scores, the server recommends safe trading assets to the user. The recommendation results are sent to the terminal, which notifies the user through a notification system. For example, a specific list of stocks may be displayed, such as "The following stocks are safe for you to buy and sell today: Y Co., Ltd." The terminal uses the notification function to provide information to the user in real time.

[1074] Examples of concrete examples and prompts

[1075] For example, if a user sends or receives an email with the content "We will discuss a new product at the next board meeting," this email is collected by the device and uploaded to the server. The server preprocesses this email and inputs it into a machine learning model, which may determine that the email poses a high insider trading risk. In this case, the risk score will be high and the related stocks will not be recommended. Instead, other low-risk stocks will be recommended to the user.

[1076] Below are some example prompts for generative AI models:

[1077] "We have detected an email containing information about a new product that the user would like to discuss at the next board meeting. Based on this information, we would like you to calculate an insider trading risk score and recommend safe stocks to trade."

[1078] In this way, the system of the present invention works in conjunction with various communication means to assist users in safely and efficiently trading stocks without unintentionally incurring legal risks.

[1079] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1080] Step 1:

[1081] Data collection

[1082] The device periodically collects data from the user's emails, chat messages, calendar events, meeting minutes, etc. For example, a Python script installed on the device uses the Microsoft Graph API to access the user's mailbox and retrieve new emails.

[1083] Input: Various user communication data (emails, chat messages, etc.)

[1084] Output: Raw data stored on the device

[1085] Step 2:

[1086] Data storage and trigger settings

[1087] The collected data is saved as a temporary file on the device's local disk. For example, meeting minutes are saved as "meeting_notes.txt." The device also uploads data at a specified time (e.g., midnight) or when a specific event (e.g., receiving an email) triggers the upload.

[1088] Input: Raw data stored on the device

[1089] Output: Prepared upload data

[1090] Step 3:

[1091] Data upload

[1092] The data is encrypted and uploaded to the server using the HTTPS protocol: the script on the device sends the collected data to the server using HTTP POST requests.

[1093] Input: Prepared upload data

[1094] Output: Data sent to the server

[1095] Step 4:

[1096] Data reception

[1097] The server receives the data sent from the device. For example, a Flask-based web server receives a POST request and saves the data in the appropriate folder (e.g., " / data / incoming").

[1098] Input: Data sent from the terminal

[1099] Output: Temporary data on the server

[1100] Step 5:

[1101] Data Preprocessing

[1102] The server performs preprocessing on the received data, including standardization, tokenization, and stop word removal. Specifically, it uses Python's Beautiful Soup library to remove HTML tags, and the NLTK library to tokenize each word and remove frequent words. For example, it removes HTML tags from the content of "meeting_notes.txt" and converts it into pure text.

[1103] Input: Temporary data on the server

[1104] Output: Preprocessed data

[1105] Step 6:

[1106] Risk Assessment

[1107] Based on the preprocessed data, the server uses machine learning models to perform risk assessments. For example, TensorFlow can be used to calculate a risk score that indicates the likelihood of insider trading. For example, minutes of a meeting about a new product might be assigned a high risk score (e.g., 90).

[1108] Input: Preprocessed data

[1109] Output: Risk score

[1110] Step 7:

[1111] Risk score aggregation

[1112] The server aggregates the risk scores obtained for each data. For example, if "Company A's risk score is 90" and "Company B's risk score is 20," then "Company B" will be included in the recommendation list.

[1113] Input: Risk Score

[1114] Output: Aggregated risk level

[1115] Step 8:

[1116] Stock Recommendations

[1117] The server recommends safe trading assets to users based on the aggregated risk level. Specifically, it generates a recommended list of assets with low risk levels.

[1118] Input: Aggregated risk level

[1119] Output: Recommended stock list

[1120] Step 9:

[1121] Notification of recommendation results

[1122] The server sends the recommendation results to the terminal, which then notifies the user. Specifically, the terminal's notification system is used to provide real-time information to the user. For example, it may notify the user that "This stock is safe to trade today: Co. B."

[1123] Input: Recommended stock list

[1124] Output: User notification

[1125] By following the steps above, users can safely trade stocks while minimizing the risk of insider trading.

[1126] (Application example 1)

[1127] 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."

[1128] Currently, it is extremely important to properly manage insider trading risks in stock trading. However, existing systems inadequately assess the risk of stocks traded by users, and manual confirmation work requires a great deal of effort. Furthermore, even when safe stocks are recommended, the procedures for immediate purchase are cumbersome, hindering smooth trading. Therefore, the present invention aims to provide a system that supports users in safely and efficiently trading stocks, enabling risk assessment and immediate purchase.

[1129] 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.

[1130] In this invention, the server includes means for evaluating a risk score associated with insider trading using a machine learning model based on the preprocessed data, means for aggregating multiple risk scores to determine a final risk level, and means for instantly purchasing safe stocks in cooperation with the user's electronic payment platform based on the aggregated risk level, thereby enabling the user to quickly purchase safe trading stocks.

[1131] A "terminal" is an electronic device used by a user, such as a smartphone or PC, that collects and notifies data.

[1132] A "server" is a computer system that receives, processes, and evaluates data sent from a terminal.

[1133] "Communication tools" are tools used by users, such as email, chat messages, and calendar events.

[1134] "Materials" refers to document data related to a transaction, such as a user's electronic documents or meeting minutes.

[1135] "Data collection" is the process by which the device periodically collects data from the user's communication tools and materials.

[1136] "Data preprocessing" is the process by which the server standardizes, tokenizes, removes stop words, etc., data sent from the terminal.

[1137] A "machine learning model" is a model that is trained based on past insider trading data and performs risk assessments on new data.

[1138] The "risk score" is a numerical representation of insider trading risk calculated using a machine learning model.

[1139] The "risk level" is an index that indicates the safety of a trading stock and is determined by aggregating multiple risk scores.

[1140] "Trading assets" are financial assets such as stocks and securities that users buy and sell.

[1141] "Recommendation" refers to the act of the server suggesting trading assets to a user based on risk level.

[1142] "Notification" is the process by which the terminal notifies the user of the recommendation results received from the server.

[1143] An "electronic payment platform" is an online payment system used to purchase and sell trading assets.

[1144] "Linkage" is the process by which the server communicates data with the electronic payment platform to realize the transaction.

[1145] The system of the present invention involves a terminal and a server working together to evaluate the risk of insider trading in stock trading and recommend safe stocks to users. The components of this system and their operation are described below.

[1146] Data collection

[1147] The device periodically collects data from the communication tools (emails, chat messages, calendar events, etc.) and documents used by the user. This collected data is temporarily stored on the device.

[1148] Uploading data

[1149] At a specified time or trigger, the collected data is consolidated and uploaded to the server by the device, and the uploaded data is securely transmitted to the server via the Internet.

[1150] Data Preprocessing

[1151] The server preprocesses the received data, which includes standardizing the text data (removing HTML tags, etc.), tokenizing (splitting words), removing stop words (removing non-important words), etc. Spacy is used as the natural language processing library.

[1152] Risk Assessment

[1153] Based on the preprocessed text data, the server uses a generative AI model (machine learning model) to calculate a risk score. This model is trained on past insider trading data and assesses the insider trading risk of new data.

[1154] Determining the risk level

[1155] The server aggregates multiple risk scores and determines a final risk level, which evaluates the overall insider trading risk for each stock.

[1156] Stock recommendation and electronic payment platform integration

[1157] The server recommends safe trading assets to users based on the aggregated risk level, and also provides a means for users to instantly purchase safe stocks in conjunction with their electronic payment platform, allowing users to quickly purchase the recommended stocks.

[1158] Notification of recommendation results

[1159] Finally, the server sends the recommendation results to the device, which then notifies the user via a smartphone application or a PC notification system.

[1160] Specific examples

[1161] For example, if a user sends or receives an email with the content "At the next meeting, we will discuss a new product from Company X," this email is collected by the device and uploaded to the server. The server preprocesses this email and inputs it into a machine learning model to determine that it poses a high risk of insider trading. In this case, the risk score will be high, and the related Company X's stock will be excluded as a non-recommended stock. Meanwhile, other low-risk stocks will be recommended to the user.

[1162] An example of a prompt sentence to input to the generative AI model is as follows:

[1163] At the next meeting, we will discuss Company X's new product.

[1164] Yu

[1165] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1166] Step 1:

[1167] The device collects data from the user's communication tools and documents. Input data includes emails, chat messages, calendar events, etc., and periodically scans them to extract important information and temporarily store it on the device. The output is a list of the collected data.

[1168] Step 2:

[1169] The device uploads the collected data to the server. Here, the data is encrypted and securely sent to the server at regular intervals or when a specific trigger (user-specified operation or detection of new data) is triggered. The input is the data stored on the device, and the output is the data received by the server.

[1170] Step 3:

[1171] The server preprocesses the received data. Specifically, it standardizes, tokenizes, and removes stop words. Spacy is used as a natural language processing library, and the input data is text data, and the output is preprocessed, cleaned text data.

[1172] Step 4:

[1173] The server evaluates a risk score associated with insider trading using a generative AI model based on the preprocessed data. The machine learning model is trained on historical insider trading data and takes the preprocessed text data as input. The output is a calculated insider trading risk score.

[1174] Step 5:

[1175] The server aggregates multiple risk scores and determines the final risk level. Here, the risk scores for each data are integrated to perform an overall risk assessment. The input is the individual risk scores, and the output is the integrated final risk level.

[1176] Step 6:

[1177] The server recommends safe trading assets to users based on their risk level. Depending on the risk level, stocks deemed safe are selected and a list is generated. The input is the final risk level, and the output is a list of recommended safe trading assets.

[1178] Step 7:

[1179] The server instantly purchases safe stocks based on the aggregated risk level in conjunction with the user's electronic payment platform. The server uses the user's account information and payment system to automate the purchase process for the recommended stocks. The input is a list of recommended trading assets, and the output is confirmation of the completion of the purchase process.

[1180] Step 8:

[1181] The terminal notifies the user of the recommendation results and purchase completion notification. Specifically, the recommended stocks and purchase results are notified to the user via a smartphone application or a PC notification system. The input is the notification content sent from the server, and the output is what is displayed to the user.

[1182] 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.

[1183] The system of the present invention operates in cooperation with a terminal and a server to support users in safely trading stocks. This system also incorporates an emotion engine that recognizes the user's emotions, and the results can be used to assess the risk of insider trading. Specific embodiments are described below.

[1184] Data collection and upload

[1185] The device collects data such as the user's emails, chat messages, voice data, calendar events, and meeting minutes. This collection process is automated and runs periodically. For example, the device accesses the user's mailbox every hour to retrieve new emails and voice messages. Calendar and meeting minutes data is also retrieved in the same way. This data is temporarily stored in the device's local storage and periodically uploaded to the server.

[1186] Data Preprocessing

[1187] When the server receives data sent from a terminal, it first preprocesses the data. Preprocessing includes cleaning the text data. Specifically, HTML tags are removed from emails and chat messages to convert them into pure text. Natural language processing is then used to tokenize the data and remove meaningless words (stop words). In the case of audio data, speech recognition technology is used to convert it into text.

[1188] Emotion analysis

[1189] The server uses the preprocessed text and voice data to perform emotion analysis using an emotion engine. The emotion engine recognizes the user's emotions (happiness, anger, sadness, etc.) from the text data. It also analyzes the tone and intonation of the voice data to recognize similar emotions. This emotion data is then incorporated into the risk assessment process.

[1190] Risk Assessment

[1191] Once preprocessing and sentiment analysis are complete, the server uses a machine learning model to perform risk assessment on the data. The model analyzes the preprocessed text data, audio data, and sentiment data together to calculate a risk score related to insider trading. This score indicates the risk level of each data.

[1192] Risk score aggregation

[1193] The risk scores calculated for each data item are aggregated by the server, and the final risk level is determined. Specifically, the risk scores obtained for each data item are aggregated using statistical methods to calculate the overall risk level for each stock. This allows an evaluation of whether a particular stock can be traded.

[1194] Stock recommendations and notifications

[1195] The server recommends safe stocks to the user based on the aggregated risk level. This involves the process of selecting and listing stocks with low risk scores. The selected stocks are recommended to the user and ultimately sent to the user's device. The user's device receives the recommendation results and notifies the user. The device displays a pop-up notification or email with a "list of stocks that are safe to buy and sell today," providing the user with information to trade safely. This notification allows the user to proceed with trading with peace of mind.

[1196] Specific examples

[1197] For example, if a user sends and receives an email stating, "At the next board meeting, we will discuss a new product from Company X," along with a voice message expressing joy, this data is collected by the device and uploaded to the server. The server preprocesses this email and voice data and uses an emotion engine to recognize that the user is expressing joy. The server then inputs this data into a machine learning model and determines that the insider trading risk is high (high risk score). In this case, shares of Company X, which has a high risk level, are not recommended to the user, and instead other low-risk stocks are recommended to the user.

[1198] This system allows users to trade stocks safely without worrying about the risk of insider trading, and reduces legal risks for the company as a whole. The addition of an emotion engine enables risk assessment that takes user emotions into account, resulting in more accurate recommendations. Each component of this system is built by combining existing technologies, making it a widely applicable and feasible solution.

[1199] The processing flow will be explained below.

[1200] Step 1:

[1201] The device collects data such as the user's emails, chat messages, voice data, calendar events, and meeting minutes. This collection process is automated, so for example, the device accesses the user's mailbox every 30 minutes to retrieve new emails and newly recorded voice data. Calendar information and meeting minutes data are also collected in the same way.

[1202] Step 2:

[1203] The device uploads the collected data to the server. The collected data is temporarily stored in the device's local storage and then transferred to the server based on a set schedule. At this time, the data is encrypted to ensure security.

[1204] Step 3:

[1205] The server preprocesses the data it receives. Preprocessing includes cleaning the text data. For example, HTML tags are removed from emails and chat messages to convert them into pure text. Natural language processing technology is then used to tokenize the data and remove meaningless words (stop words). In the case of audio data, speech recognition technology is used to convert it into text.

[1206] Step 4:

[1207] The server uses an emotion engine to analyze the user's emotions from the preprocessed text and voice data. The emotion engine uses, for example, natural language processing (NLP) techniques to recognize the user's emotions (happiness, anger, sadness, etc.) from the text data. It also analyzes the tone and intonation of the voice data to recognize similar emotions. The recognized emotion data is then incorporated into the subsequent risk assessment process.

[1208] Step 5:

[1209] The server uses the sentiment-analyzed data to perform risk assessment using a machine learning model. The machine learning model is a trained model that calculates a risk score related to insider trading. The model takes the preprocessed text data, audio data, and sentiment data as input and outputs a numerical risk level for each data.

[1210] Step 6:

[1211] The server aggregates multiple risk scores to determine the final risk level. Specifically, the risk scores obtained for each data point are aggregated using statistical methods such as weighted averaging to calculate an overall risk level for each stock. This is used to evaluate whether a particular stock can be traded.

[1212] Step 7:

[1213] The server recommends safe trading stocks to the user based on the risk level. Stocks with low risk scores are selected and listed. This recommendation list is sent from the server to the user's terminal.

[1214] Step 8:

[1215] The terminal notifies the user of the recommendation results sent from the server. Specifically, the terminal displays a "list of stocks that are safe to buy and sell today" via a pop-up notification or email, providing the user with information to trade safely. This notification allows the user to trade with peace of mind.

[1216] Example 2

[1217] 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."

[1218] The challenge is to provide a system that reduces the risk of insider trading and supports safe stock trading. Conventional systems do not take user sentiment or unstructured data into account, which can result in low accuracy in risk assessment and the possibility of incorrect trade recommendations. Furthermore, data preprocessing and risk assessment require a significant amount of time and effort, so automation and improved accuracy are required.

[1219] 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.

[1220] In this invention, the server includes means for preprocessing received data, means for analyzing user emotions based on the preprocessed data using an emotion recognition engine, and means for evaluating a risk score related to insider trading using a machine learning model based on the emotion-analyzed data. This enables highly accurate risk assessment that takes user emotions and unstructured data into consideration, and the recommendation of safe stocks to trade.

[1221] "Terminal" means a computing device used by a user, primarily a hardware device or software application for data collection, processing, and notification.

[1222] "Server" refers to a computer system that receives data sent from terminals and performs centralized processing such as pre-processing, sentiment analysis, risk assessment, and stock recommendations.

[1223] An "emotion recognition engine" is a machine learning model or algorithm that analyzes a user's emotions from text and voice data and identifies the type and intensity of those emotions.

[1224] A "machine learning model" is an algorithm that learns patterns and trends based on past data and makes predictions and classifications for new data, and is used particularly for risk assessment in the present invention.

[1225] A "risk score" is a numerical representation of the degree of risk associated with insider trading, and is an index calculated for each data point.

[1226] "Preprocessing" is the process of applying initial processing to received data, including text standardization, tokenization, stop word removal, and converting audio data to text.

[1227] "Communication Tools" means software or platforms that allow users to send and receive messages and make voice and video calls, such as email clients and chat applications.

[1228] An "information source" is a source of digital content that a user accesses, such as a mail server, chat application, or calendar system.

[1229] "Uploading" is the process by which a device sends collected data to a server, using a secure transfer protocol such as SFTP or HTTPS.

[1230] "Stock recommendation" is the process by which the server suggests safe trading stocks to users based on risk assessment.

[1231] "Notification" refers to the process by which the device notifies the user of the recommendation results sent from the server, and includes providing information via pop-up notifications or emails.

[1232] System Overview

[1233] The system of the present invention, in which a terminal and a server work in cooperation with each other, supports users in safely trading stocks. This system also incorporates an emotion engine that recognizes the user's emotions, and the results can be used to assess the risk of insider trading. Specific embodiments are described below.

[1234] Data collection by terminal

[1235] The device collects data such as the user's emails, chat messages, voice data, calendar events, and meeting minutes. This collection process is automated and runs periodically. For example, the device accesses the user's mailbox every hour to retrieve new emails and voice messages. This is done using the IMAP protocol. Additionally, calendar events are retrieved using the Google Calendar API. This data is temporarily stored in the device's local storage and then securely uploaded to the server using, for example, SFTP or HTTPS.

[1236] Data preprocessing by the server

[1237] When the server receives data sent from the device, it first preprocesses the data. HTML tags are removed from emails and chat messages and converted into pure text. This is done using Python's BeautifulSoup. Natural language processing techniques are then used to tokenize the text data and remove stop words. Voice data is converted into text using the Google Speech-to-Text API.

[1238] Emotion analysis by server

[1239] Based on the pre-processed text and audio data, the server performs sentiment analysis using an emotion engine. The text data is input into a sentiment analysis model (e.g., BERT) and an emotion label is output. The audio data is analyzed for tone and intonation to identify emotions. This emotion data is used in the subsequent risk assessment process.

[1240] Server Risk Assessment

[1241] After data preprocessing and sentiment analysis are completed, the server performs risk assessment using a machine learning model (e.g., XGBoost). It calculates an insider trading risk score based on the preprocessed text data, voice data, and sentiment data. This risk score indicates the risk level of each data.

[1242] Risk score aggregation and stock recommendations

[1243] The server uses statistical methods to compile the risk scores calculated for each data point and determine the overall risk level for each stock. Stocks that are deemed safe are sent from the server to the user's device as recommendation information. The device receives this recommendation and presents the user with a "list of stocks that are safe to buy and sell today" via a pop-up notification or email.

[1244] Specific examples

[1245] For example, if a user sends or receives an email stating, "At the next meeting, we will discuss Company A's new product," along with a voice message expressing their feelings of joy, this data is collected by the device. The device then uploads this email and voice data to the server. The server then preprocesses the data by removing HTML tags and converting it into text. The emotion engine then recognizes the user's emotion of joy. This data is then input into a machine learning model, which evaluates the risk of insider trading as high. In this case, Company A's stock, which has a high risk level, is not recommended, and instead other low-risk stocks are recommended to the user.

[1246] Prompt Sentence Examples

[1247] "Please explain in detail how the system preprocesses data collected by the device, including emails, chat messages, voice data, calendar events, and meeting minutes, and then uses an emotion engine to perform risk assessments. This ensures secure transactions."

[1248] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1249] Step 1: Data collection

[1250] The device periodically collects data such as the user's emails, chat messages, voice data, calendar events, and meeting minutes. For example, the device accesses the mail server every hour using the IMAP protocol to download new emails. It also retrieves calendar events using the Google Calendar API and saves them in local storage.

[1251] Input: Mail server, chat app, voice data source, calendar system

[1252] Output: New emails, chat messages, audio files, and calendar events stored in local storage.

[1253] Step 2: Upload data

[1254] The device uploads the collected data to a server at regular intervals, and to ensure security, the data is transferred using SFTP or HTTPS protocols.

[1255] Input: Data in local storage (emails, chat messages, audio files, calendar events)

[1256] Output: Report that data upload to server is complete

[1257] Step 3: Data Preprocessing

[1258] The server preprocesses the data received from the device. First, it uses Python's BeautifulSoup to remove HTML tags from emails and chat messages and convert them into pure text. Next, it tokenizes the text data using natural language processing techniques and removes stop words. For audio data, it converts it to text using the Google Speech-to-Text API.

[1259] Input: Uploaded emails, chat messages, audio files, calendar events

[1260] Output: Preprocessed clean text data and converted audio data

[1261] Step 4: Sentiment Analysis

[1262] The server runs an emotion engine using the preprocessed text and audio data. The emotion engine uses an emotion analysis model such as the BERT model to recognize user emotions (e.g., joy, anger, sadness) from the text data. It also analyzes tone and intonation from the audio data to recognize similar emotions.

[1263] Input: Preprocessed text data, audio data

[1264] Output: Emotional information tagged to text and audio data

[1265] Step 5: Risk assessment

[1266] The server calculates an insider trading risk score using a machine learning model (e.g., XGBoost) based on the pre-processed and sentiment-analyzed data. It integrates the pre-processed text data, voice data, and sentiment data to evaluate the risk level of each.

[1267] Input: Sentiment-analyzed text data, audio data

[1268] Output: Risk score for each data point

[1269] Step 6: Aggregate risk scores

[1270] The server then uses statistical methods to aggregate the calculated risk scores to determine the overall risk level for each security. This process involves evaluating the overall risk scores for each data point and identifying the security deemed to be the least risky.

[1271] Input: Individual Risk Score

[1272] Output: Overall risk level for each stock

[1273] Step 7: Stock Recommendation and Notification

[1274] The server recommends safe stocks to the user based on the aggregated risk level. The list of selected stocks is sent to the terminal. The terminal notifies the user via a pop-up notification or email of the "list of stocks that are safe to buy and sell today."

[1275] Input: Aggregated risk level

[1276] Output: Recommended stock notification to user device

[1277] In this way, the system of the present invention helps users to trade stocks safely.

[1278] (Application example 2)

[1279] 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."

[1280] In stock trading, there is a risk of insider trading and inappropriate trading due to users making decisions based on their emotions. Current systems perform risk assessment without taking users' emotions into account, which limits their accuracy. In addition, they lack protection against users making large transactions while in an emotional state. Therefore, there is a need for a system that performs risk assessment taking users' emotions into account and supports safe and appropriate trading.

[1281] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for converting voice data into text, means for recognizing emotions from the text data and voice data and performing risk assessment based on the emotions, and means for suspending a transaction and requesting confirmation in a high-risk situation. This enables highly accurate risk assessment that takes the user's emotions into consideration, and supports safe and appropriate transactions.

[1282] A "terminal" is a device used by a user that collects data and uploads it to a server.

[1283] The "server" is a central processing unit that receives data sent from the terminals and performs pre-processing and risk assessment.

[1284] "Communication tools" are communication means that users use on a daily basis, such as e-mail, chat applications, and voice message applications.

[1285] "Data preprocessing" refers to data processing in which the server cleans the data it receives, tokenizes it, removes stop words, and so on.

[1286] A "machine learning model" is an algorithm that is trained based on past data and used to assess risk.

[1287] "Insider trading" is the illegal act of trading stocks using inside information.

[1288] A "risk score" is a numerical representation of the risk level of a transaction, calculated using a machine learning model.

[1289] "Emotion recognition" means analyzing emotions (joy, anger, sadness, etc.) from a user's text data or voice data.

[1290] A "high-risk situation" is one in which the risk score is high and caution or restrictions on transactions are required.

[1291] "Trading products" are items such as financial products and stocks that are the subject of trading by users.

[1292] The system of the present invention evaluates insider trading risks while taking into account user sentiment and supports safe trading. Specifically, the system is composed of a terminal, a server, and a machine learning model.

[1293] Data collection and upload

[1294] The device automatically collects data from the communication tools and materials used by the user. The collected data includes emails, chat messages, voice data, calendar events, and meeting minutes. This process is performed periodically; for example, the device accesses the user's mailbox every hour to retrieve new emails. It also collects voice recordings of voice data. This data is temporarily stored in the device's local storage and periodically uploaded to the server.

[1295] Data Preprocessing

[1296] When the server receives the data sent from the device, it first preprocesses the data. This includes cleaning and standardizing the text data, tokenizing it, and removing stop words. In the case of audio data, it is converted into text using speech recognition technology. For this purpose, it is effective to use the Python SpeechRecognition library.

[1297] Emotion analysis

[1298] The server performs sentiment analysis using an emotion engine based on the preprocessed text and audio data. The emotion engine recognizes the user's emotions (happiness, anger, sadness, etc.) from the text and audio data. For text data, it uses the TextBlob library, and for audio data, it analyzes the tone and intonation of the voice.

[1299] Risk Assessment

[1300] Once preprocessing and sentiment analysis are complete, the server uses the data to perform risk assessment using a machine learning model. This model is trained on historical insider trading data and sentiment data. The machine learning model uses the scikit-learn library and applies the logistic regression algorithm.

[1301] Risk score aggregation and transaction restrictions

[1302] The risk scores calculated for each data item are aggregated by the server to determine the final risk level. If the risk level is high, the server sends a notification to the terminal requesting a pause in the transaction and requesting the user to reconfirm. This prevents users from making high-risk transactions based on emotional judgment.

[1303] Stock recommendations and notifications

[1304] The server recommends safe trading products to users based on the final risk level. This includes the process of selecting and listing low-risk trading products. The selected products are sent to the user's device and displayed via pop-up notification or email. This allows users to proceed with trading with peace of mind.

[1305] Specific examples

[1306] For example, if a user sends a voice message saying, "I want to make a big purchase right now!", the device collects this and sends it to the server, which converts the voice data into text and performs sentiment analysis. If the sentiment score is high and the transaction is deemed high risk, the server will pause the transaction and send a notification to the user requesting confirmation.

[1307] Prompt Sentence Examples

[1308] "Analyze voice messages from users expressing their desire to make a large purchase, assess the risk of electronic payment services based on that sentiment, and pause the transaction if necessary."

[1309] This allows users to avoid emotional decisions and conduct safe and appropriate transactions.

[1310] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1311] Step 1:

[1312] The device collects data from the communication tools and materials used by the user. The collection process is automated and runs periodically. Specifically, the device accesses the user's mailbox and chat application every hour to retrieve new emails and messages. The device also records the user's voice messages, which are included in the collected data. The input is the communication tool and voice data, and the output is the collected data.

[1313] Step 2:

[1314] The device uploads the collected data to the server. This is a periodic process in which data temporarily stored in the device's local storage is transferred to the server. The input is the collected data, and the output is the data transferred to the server. Specifically, the data is sent to the server using an HTTP request.

[1315] Step 3:

[1316] The server preprocesses the received data. Preprocessing includes cleaning and standardizing the text data, tokenizing it, and removing stop words. In the case of audio data, it is first converted into text using speech recognition technology. The input is the raw data uploaded to the server, and the output is the cleaned text data and the audio data converted into text. Specific operations use the SpeechRecognition and TextBlob libraries.

[1317] Step 4:

[1318] The server performs sentiment analysis based on the preprocessed text and audio data. The sentiment engine recognizes the user's sentiment from the text and audio data. The input is the preprocessed data, and the output is an emotion score. Specifically, it uses the TextBlob library to calculate the sentiment score for the text and analyzes the tone and intonation of the audio data.

[1319] Step 5:

[1320] The server performs risk assessment using a machine learning model based on the sentiment data and preprocessed data. This model is trained based on past insider trading data and sentiment data. The input is sentiment data and preprocessed data, and the output is a risk score. Specifically, it uses the scikit-learn library and applies the Logistic Regression algorithm.

[1321] Step 6:

[1322] The server aggregates the risk scores calculated for each data item and determines the final risk level. The input is multiple risk scores, and the output is the final risk level. Specifically, the scores are aggregated using a statistical aggregation method.

[1323] Step 7:

[1324] The server recommends safe trading products to the user based on the risk level. If the risk level is high, the server suspends the transaction and sends a notification to the terminal requesting confirmation. The input is the final risk level, and the output is a list of recommended trading products and a notification. Specific operations include notifying the user via a pop-up notification or email.

[1325] Step 8:

[1326] The terminal notifies the user of the recommendation results sent from the server. The input is the recommendation results and notification request from the server, and the output is a notification message to the user. Specifically, the terminal displays a "list of trading products that are safe to buy and sell today" via a pop-up notification or email.

[1327] This series of steps allows users to avoid the risks of emotional decisions and conduct safe and appropriate transactions.

[1328] 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.

[1329] 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.

[1330] 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.

[1331] 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.

[1332] FIG. 9 illustrates 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 behaviors 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.

[1333] 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.

[1334] 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).

[1335] 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.

[1336] 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."

[1337] 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.

[1338] 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).

[1339] 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.

[1340] 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.

[1341] 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.

[1342] 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.

[1343] 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.

[1344] 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.

[1345] 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.

[1346] 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.

[1347] 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.

[1348] 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.

[1349] The following is further disclosed regarding the above embodiment.

[1350] (Claim 1)

[1351] A terminal has means for collecting data from communication tools and materials used by the user;

[1352] A means for the terminal to upload the collected data to a server;

[1353] means for the server to preprocess the received data;

[1354] a means for the server to assess a risk score associated with insider trading using a machine learning model based on the preprocessed data;

[1355] a means for the server to aggregate the multiple risk scores to determine a final risk level;

[1356] A means for the server to recommend safe trading instruments to the user based on the risk level;

[1357] a means for the terminal to notify the user of the recommendation result;

[1358] A system including:

[1359] (Claim 2)

[1360] 10. The system of claim 1, wherein the data preprocessing means performs text standardization, tokenization, and stop word removal.

[1361] (Claim 3)

[1362] 10. The system of claim 1, wherein the machine learning model comprises a risk assessment model trained on historical insider trading data.

[1363] "Example 1"

[1364] (Claim 1)

[1365] A terminal has a means for collecting data from communication means and information used by the user;

[1366] A means for the terminal to upload the collected data to a server;

[1367] means for the server to preprocess the received data;

[1368] a means for the server to assess a risk score associated with market manipulation using a machine learning model based on the preprocessed data;

[1369] a means for the server to aggregate the multiple risk scores to determine a final risk level;

[1370] A means for the server to recommend safe trading assets to users based on their risk levels;

[1371] a means for the terminal to notify the user of the recommendation result;

[1372] A system including:

[1373] (Claim 2)

[1374] 10. The system of claim 1, wherein the data preprocessing means standardizes, tokenizes, and removes frequently used words from the text.

[1375] (Claim 3)

[1376] 10. The system of claim 1, wherein the machine learning model comprises a risk assessment model trained on historical market manipulation data.

[1377] "Application Example 1"

[1378] (Claim 1)

[1379] A terminal has means for collecting data from communication tools and materials used by the user;

[1380] A means for the terminal to upload the collected data to a server;

[1381] means for the server to preprocess the received data;

[1382] a means for the server to assess a risk score associated with insider trading using a machine learning model based on the preprocessed data;

[1383] a means for the server to aggregate the multiple risk scores to determine a final risk level;

[1384] A means for the server to recommend safe trading assets to users based on their risk levels;

[1385] a means for the server to instantly purchase the safe stocks in cooperation with the user's electronic payment platform based on the aggregated risk level;

[1386] a means for the terminal to notify the user of the recommendation result;

[1387] A system including:

[1388] (Claim 2)

[1389] 10. The system of claim 1, wherein the data preprocessing means performs text standardization, tokenization, and stop word removal.

[1390] (Claim 3)

[1391] 10. The system of claim 1, wherein the machine learning model comprises a risk assessment model trained on historical insider trading data.

[1392] "Example 2: Combining Emotion Engines"

[1393] (Claim 1)

[1394] A means for the terminal to collect data from communication tools and information sources used by the user;

[1395] A means for the terminal to upload the collected data to a server;

[1396] means for the server to preprocess the received data;

[1397] A means for the server to analyze the user's emotions using an emotion recognition engine based on the preprocessed data;

[1398] a means for the server to assess a risk score related to insider trading using a machine learning model based on the sentiment analyzed data;

[1399] a means for the server to aggregate the multiple risk scores to determine a final risk level;

[1400] A means for the server to recommend safe trading instruments to the user based on the risk level;

[1401] a means for the terminal to notify the user of the recommendation result;

[1402] A system including:

[1403] (Claim 2)

[1404] 10. The system of claim 1, wherein the data preprocessing means performs text standardization, tokenization, stop word removal, and text conversion of the audio data.

[1405] (Claim 3)

[1406] 10. The system of claim 1, wherein the machine learning model comprises a risk assessment model trained on historical insider trading data and sentiment data.

[1407] "Application example 2 when combining emotion engines"

[1408] (Claim 1)

[1409] A terminal has means for collecting data from communication tools and materials used by the user;

[1410] A means for the terminal to upload the collected data to a server;

[1411] means for the server to preprocess the received data;

[1412] a means for the server to assess a risk score associated with insider trading using a machine learning model based on the preprocessed data;

[1413] a means for the server to aggregate the multiple risk scores to determine a final risk level;

[1414] A means for the server to recommend safe trading products to the user based on the risk level;

[1415] a means for the terminal to notify the user of the recommendation result;

[1416] A terminal has means for collecting voice data and transmitting the voice data to a server;

[1417] a means for converting the voice data into text by the server;

[1418] a server that recognizes emotions from text data and voice data and performs risk assessment based on the emotions;

[1419] A means to suspend transactions and require confirmation in high-risk situations;

[1420] A system including:

[1421] (Claim 2)

[1422] 10. The system of claim 1, wherein the data preprocessing means performs text standardization, tokenization, and stop word removal.

[1423] (Claim 3)

[1424] 10. The system of claim 1, wherein the machine learning model comprises a risk assessment model trained on historical insider trading data and sentiment data. [Explanation of symbols]

[1425] 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 terminal has means for collecting data from communication tools and materials used by the user; A means for the terminal to upload the collected data to a server; means for the server to preprocess the received data; a means for the server to assess a risk score associated with insider trading using a machine learning model based on the preprocessed data; a means for the server to aggregate the multiple risk scores to determine a final risk level; A means for the server to recommend safe trading products to the user based on the risk level; a means for the terminal to notify the user of the recommendation result; A system including:

2. The system of claim 1 , wherein the data preprocessing means performs text standardization, tokenization, and stop word removal.

3. The system of claim 1 , wherein the machine learning model comprises a risk assessment model trained on historical insider trading data.

Images