Machine learning device, data processing device, inference device, machine learning method, data processing method, and inference method
Patent Information
- Application Number
- JP2021213643
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2021-12-28
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2041-12-28
AI Technical Summary
Existing systems fail to provide appropriate information based on behavioral factor indexes inherent in users, as these indexes are not input as personal information, leading to inadequate targeting of user actions.
A machine learning device that learns correlations between person data and behavioral factor indexes, allowing for the estimation of these indexes without requiring personal information input, using a machine learning unit to train a model that can assign behavioral factor indexes to person data.
Enables the accurate assignment of behavioral factor indexes to users, facilitating personalized services without the need for explicit personal information input, improving the relevance of information provision.
Smart Images

Figure 00000000_0000_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a machine learning device, a data processing device, an inference device, a machine learning method, a data processing method, and an inference method.
Background Art
[0002] Conventionally, by using personal information of a user and providing information tailored to that user, the convenience of the user and the usefulness of the information provider have been improved. For example, Patent Document 1 and Patent Document 2 disclose a system that distributes advertisement information to a user's terminal based on personal information including the user's name, gender, age, and the like.
Prior Art Documents
Patent Documents
[0003]
Patent Document 1
特許文献۲
Summary of the Invention
Problems to be Solved by the Invention
[0004] In the systems disclosed in Patent Document 1 and Patent Document 2, since it is assumed that personal information is input by each user, it has been impossible to provide appropriate information to users who have not input personal information. Further, when a user takes some action, it is considered that a behavior factor index (for example, sensitivity, etc.) indicating a factor that motivates the action is inherent in the user. However, in the systems disclosed in Patent Document 1 and Patent Document 2, since such a behavior factor index is not input as personal information, it has been impossible to provide appropriate information according to the behavior factor index inherent in each user.
[0005] The present invention has been made in view of the above-mentioned problems, and aims to provide a machine learning device, a data processing device, an inference device, a machine learning method, a data processing method, and an inference method for appropriately understanding the factors that motivate a person to take action based on the person's actions. [Means for solving the problem]
[0006] To achieve the above objective, a machine learning apparatus according to one aspect of the present invention is: A machine learning unit inputs multiple sets of training data into a learning model, which consist of person data recording a person's actions based on one or more perspectives and behavioral factor indicators indicating the factors that motivate the person to perform the actions, thereby training the learning model to learn the correlation between the person data and the behavioral factor indicators through machine learning. The system includes a trained model storage unit that stores the trained model, which has learned the correlation relationship by the machine learning unit. [Effects of the Invention]
[0007] According to one aspect of the present invention, a machine learning device can provide a learning model capable of estimating behavioral factor indicators that show the factors motivating a person to take action, based on person data that records a person's actions from one or more perspectives. Therefore, by using this learning model, for example, behavioral factor indicators inherent in a person (user) can be appropriately assigned (inferred) without requiring the person (user) to input personal information.
[0008] Other issues, configurations, and effects will be clarified in the embodiments for carrying out the invention described later. It can be done. [Brief explanation of the drawing]
[0009] [Figure 1] This is an overall configuration diagram showing an example of a person information management system 1. [Figure 2] This is a hardware configuration diagram showing an example of the Computer 900. [Figure 3]This is a data structure diagram showing an example of a person database 20. [Figure 4] This is a block diagram showing an example of machine learning device 3. [Figure 5] This figure shows an example of judgment rule data 310. [Figure 6] This figure shows an example of training data 11 and a training model 12. [Figure 7] This is a functional diagram illustrating an example of the machine learning unit 301. [Figure 8] An example of data processing device 4 is shown below. [Figure 9] This is a functional diagram illustrating an example of a data processing device 4. [Figure 10] This flowchart shows an example of a machine learning method using machine learning device 3. [Figure 11] This flowchart (a continuation of Figure 10) shows an example of a machine learning method using machine learning device 3. [Figure 12] This flowchart shows an example of a data processing method by the data processing device 4. [Modes for carrying out the invention]
[0010] Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings. In the following, the scope necessary for explaining how to achieve the objectives of the present invention will be schematically shown, and the scope necessary for explaining the relevant parts of the present invention will be mainly explained, with any parts that are omitted from explanation being based on prior art.
[0011] (Person Information Management System 1) Figure 1 is an overall diagram showing an example of the Person Information Management System 1. The Person Information Management System 1 functions as a system that can provide each person's behavioral factor indicators to various services (product planning, product development, production management, procurement and ordering management, sales strategy formulation, marketing, advertising, etc.) by assigning behavioral factor indicators (tags) that indicate the factors that motivate a person to take a particular action to person data that records the actions of that person.
[0012] Person data is data that records a person's actions based on one or more perspectives. Examples of perspectives when recording a person's actions include purchase behavior when a person purchases goods or services at a store or an e-commerce site, web browsing behavior when a person browses a website, web search behavior when a person searches on a website, movement behavior when a person moves by any means of transportation (walking, train, car, etc.), questionnaire response behavior when a person responds to a questionnaire on a website or in print, and the like. Therefore, person data includes, for example, a purchase history that records purchase behavior, a web browsing history that records web browsing behavior, a web search history that records web search behavior, a movement history that records movement behavior, a questionnaire response history that records questionnaire response behavior, and the like.
[0013] An action factor indicator indicates a factor that motivates a person to perform a specific action. Action factor indicators have a plurality of types classified, for example, by sensibility, preference, attributes, and the like. Examples of sensibility and preference include health orientation, organic, particular about origin, price (saving) orientation, high-class orientation, safety and security, enhancing beauty sensitivity, liking cleanliness, Japanese cuisine preference, Western cuisine preference, liking sweet things (sweet tooth), liking spicy things (spicy tooth), liking strong flavors, liking sour flavors, liking carbonated drinks, indoor type, outdoor type, and the like. Examples of attributes include those related to demographics, lifestyle, etc., such as age, gender, single, married, having children, owning a house, renting, owning a car, keeping a pet, on a diet, and the like.
[0014] The behavioral factor index is assigned based on the value of a variable for person data, and the value of the variable is assigned for each type of behavioral factor index. In the present embodiment, for the sake of explanation, mainly taking the case where the behavioral factor index of "health orientation" is defined by a two-class classification (binary classification of 0 or 1), the value of the variable of the behavioral factor index of "health orientation" is assigned as "0: not health-oriented" and "1: health-oriented". Note that the behavioral factor index may be defined by a multi-class classification (multi-value classification). For example, when it is defined by a three-class classification, the value of the variable of the behavioral factor index of "health orientation" may be assigned as "-1: not health-oriented", "0: neither", and "1: health-oriented". Also, the definition of the behavioral factor index may be different for each type of behavioral factor index.
[0015] The person information management system 1 includes, as its main components, a database device 2, a machine learning device 3, a data processing device 4, and an operator terminal device 5. Each of the devices 2 to 5 is composed of, for example, a general-purpose or dedicated computer (see FIG. 2 described later), and is connected to a wired or wireless network 6 so as to be able to transmit and receive various data to and from each other. Note that the number of each of the devices 2 to 5 and the connection configuration of the network 6 are not limited to the example of FIG. 1 and may be changed as appropriate.
[0016] The database device 2 is linked with an external system (not shown) that records the daily actions of each person as an action history based on one or more viewpoints, and includes a person database 20 capable of registering the action history based on one or more viewpoints as person data for each person. The external system is composed of, for example, a point-of-sale information management system (POS system), a computer operation log collection system, a navigation system, a questionnaire system, etc. The external system records, as an action history, a purchase history, a web browsing history, a web search history, a movement history, a questionnaire response history, etc. according to the daily actions of each person, and provides them to the database device 2 in response to a request from the database device 2, for example.
[0017] The machine learning device 3 operates as the main component of the machine learning training phase, and performs machine learning on the training model 12 by using the person data registered in the person database 20 as training data 11. The trained training model 12 is provided to the data processing device 4 via the network 6, recording media, etc. Supervised learning is employed as the machine learning method.
[0018] The data processing device 4 operates as the main component of the machine learning inference phase, and uses the trained learning model 12 generated by the machine learning device 3 to assign behavioral factor indicators to the person data, and provides the assignment results (in this embodiment, the values of variables for each type of behavioral factor indicator) to the database device 2, the worker terminal device 5, etc.
[0019] The worker terminal device 5 is a terminal device used by the worker 10 when performing machine learning on the learning model 12, specifically when annotating the training data 11. The worker terminal device 5 accepts various input operations via a display screen such as an application program or a web browser, and also displays various information (e.g., person data, behavioral factor indicators, etc.) via the display screen.
[0020] (Computer 900) Figure 2 is a hardware configuration diagram showing an example of computer 900. Each of the devices 2 to 5 of the person information management system 1 is composed of a general-purpose or dedicated computer 900.
[0021] As shown in Figure 2, the computer 900 has the following main components: a bus 910, a processor 912, memory 914, an input device 916, an output device 917, a display device 918, a storage device 920, a communication interface unit 922, an external device interface unit 924, an I / O device interface unit 926, and a media input / output unit. It includes 928. Note that the above components may be omitted as appropriate depending on the intended use of the computer 900.
[0022] The processor 912 consists of one or more arithmetic processing units (CPU (Central Processing Unit), MPU (Micro-processing unit), DSP (digital signal processor), GPU (Graphics Processing Unit), etc.) and operates as a control unit that oversees the entire computer 900. The memory 914 stores various data and programs 930 and consists of volatile memory (DRAM, SRAM, etc.) that functions as main memory, and non-volatile memory (ROM), flash memory, etc.
[0023] The input device 916 consists of, for example, a keyboard, mouse, numeric keypad, or electronic pen, and functions as an input unit. The output device 917 consists of, for example, a sound (voice) output device or a vibration device, and functions as an output unit. The display device 918 consists of, for example, a liquid crystal display, an organic EL display, electronic paper, or a projector, and functions as an output unit. The input device 916 and the display device 918 may be configured as an integrated unit, such as a touch panel display. The storage device 920 consists of, for example, an HDD (Hard Disk Drive) or an SSD (Solid State Drive), and functions as a storage unit. The storage device 920 stores various data necessary for the execution of the operating system and the program 930.
[0024] The communication I / F unit 922 is connected by wire or wireless to a network 940 such as the Internet or an intranet (which may be the same as network 6 in Figure 1) and functions as a communication unit that sends and receives data with other computers according to a predetermined communication standard. The external device I / F unit 924 is connected by wire or wireless to external devices 950 such as cameras, printers, scanners, and reader / writers and functions as a communication unit that sends and receives data with external devices 950 according to a predetermined communication standard. The I / O device I / F unit 926 is connected to I / O devices 960 such as various sensors and actuators and functions as a communication unit that sends and receives various signals and data with the I / O devices 960, for example, detection signals from sensors and control signals to actuators. The media input / output unit 928 is composed of a drive device such as a DVD (Digital Versatile Disc) drive or a CD (Compact Disc) drive and reads and writes data to media (non-temporary storage medium) 970 such as DVDs and CDs.
[0025] In the computer 900 having the above configuration, the processor 912 calls and executes the program 930 stored in the storage device 920 in the memory 914, and controls various parts of the computer 900 via the bus 910. The program 930 may also be stored in the memory 914 instead of the storage device 920. The program 930 may be recorded on the media 970 in an installable or executable file format and provided to the computer 900 via the media input / output unit 928. The program 930 may also be provided to the computer 900 by downloading it via the network 940 through the communication interface unit 922. Furthermore, the computer 900 may implement various functions realized by the processor 912 executing the program 930 using hardware such as an FPGA (field-programmable gate array) or ASIC (application-specific integrated circuit).
[0026] Computer 900 is an electronic device of any form, consisting of, for example, a stationary computer or a portable computer. Computer 900 may be a client computer, a server computer, or a cloud computer.
[0027] (Person Database 20) Figure 3 is a data structure diagram showing an example of a person database 20. The person database 20 is a database for managing person data based on person identification information (user ID) assigned to each person. The person database 20 consists of, for example, a purchase history table, a web browsing history table, a web search history table, a movement history table, a survey response history table, a behavioral factor indicator table, product master information, store master information, website master information, and map information.
[0028] The purchase history table has multiple records for recording purchasing behavior, and each record registers the date and time, product, price, store (which may be an e-commerce site), etc. A product ID is registered for each product, and by linking this product ID to the product master information, the product attributes (product name, product category, etc.) registered in the product master information are associated with it. A store ID is registered for each store, and by linking this store ID to the store master information, the store attributes (store name, store category, etc.) registered in the store master information are associated with it. Products may also be assigned product indicators (behavioral characteristic indicators) corresponding to behavioral factor indicators, for example, they may be registered in the product master information as part of the product attributes. Similarly, stores may be assigned store indicators (behavioral characteristic indicators) corresponding to behavioral factor indicators, for example, they may be registered in the store master information as part of the store attributes.
[0029] The web browsing history table has multiple records for recording web browsing behavior, with each record containing information such as the date and time, website, and browsing time. The web search history table has multiple records for recording web search behavior, with each record containing information such as the date and time and search terms. Each website has a URL for accessing it, and this URL links it to the website master information, thereby associating it with website attributes (administrator name, website category, etc.) registered in the website master information. Websites may also be assigned website indicators (behavioral characteristic indicators) corresponding to behavioral factor indicators, which may, for example, be registered in the website master information as part of the website attributes.
[0030] The movement history table has multiple records for recording movement behavior, with each record registering the date and time, location, means of transportation, etc. The location records include, for example, the place or region where a user stayed for a predetermined period of time or longer. These locations or regions are linked to map information, and the location attributes (location name, location category, etc.) and region attributes (region name, region category, etc.) registered in the map information are associated with them. Locations and regions may also be assigned location indicators (behavioral characteristic indicators) and region indicators (behavioral characteristic indicators) corresponding to behavioral factor indicators, and these may be registered in the map information as part of the location or region attributes, for example.
[0031] The survey response history table contains multiple records for recording survey response behavior, with each record including the date and time, the question asked, and the answer given.
[0032] The behavioral factor index table has multiple records for storing the behavioral factor indexes assigned by the data processing device 4, and each record registers the date and time, the value of the behavioral factor index variable, etc. In the example in Figure 3, three fields corresponding to three behavioral factor indexes, "health-conscious," "sweet tooth," and "car ownership," are shown, but similar fields are provided for other behavioral factor indexes as well.
[0033] In the person database 20, the information in each table is linked by the user ID. Person data, which records a person's behavior based on multiple perspectives, and behavioral factor indicators are managed for each person. The information registered in the person database 20 is configured to allow various statistical processing to be performed. Person data may also be recorded in association with characteristic data (in this embodiment, product, store, website, location, region) to which behavioral characteristic indicators (in this embodiment, product indicator, store indicator, website indicator, location indicator, region indicator) corresponding to behavioral factor indicators are attached, which are indicators that show the characteristics of behavior. In addition to the above, behavioral characteristic indicators may also be attached to characteristics such as time of day, day of the week, and season. Furthermore, the data structure of the person database 20 may be changed as appropriate, and data other than those mentioned above may be registered, or some of the above data may be omitted.
[0034] (Machine learning device 3) Figure 4 is a block diagram showing an example of a machine learning device 3. The machine learning device 3 comprises a control unit 30 composed of a processor, a storage unit 31 composed of an HDD, SSD, memory, etc., a communication unit 32 which is a communication interface with the network 6, an input unit 33 composed of a keyboard, mouse, etc., and a display unit 34 composed of a display, etc. Note that the input unit 33 and the display unit 34 may be omitted.
[0035] The memory unit 31 stores the training data 11, the training model 12, the decision rule data 310, and the machine learning program 311, as well as the operating system, other programs, various types of data, etc. The memory unit 31 functions as a training data storage unit that stores the training data 11, and a trained model storage unit that stores the training model 12.
[0036] Figure 5 shows an example of judgment rule data 310. The judgment rule data 310 contains judgment rules for determining whether a person conforms to a behavioral factor indicator, with each type of behavioral factor indicator registered. The judgment rule data 310 is information that can be edited by the worker 10 via the worker terminal device 5.
[0037] The judgment rules are defined for individual data based on at least some of the perspectives, and are defined by appropriately combining various numerical calculations with comparisons, proportions, frequencies, periods, categories, behavioral characteristic indicators (product indicators, store indicators, website indicators, location indicators, regional indicators), etc. Furthermore, the judgment rules are defined for each type of behavioral factor indicator, and in the example in Figure 5, the judgment rules for three behavioral factor indicators, "health-conscious," "sweet tooth," and "car owner," are defined for "purchase history" based on the perspective of "purchase behavior." For example, in the judgment rule for the behavioral factor indicator "health-conscious," the purchase amount for each user when purchasing products assigned the "health food" product category or the "health-conscious" product indicator in the past year is aggregated, and if the aggregated purchase amount is X yen or more, the behavioral factor indicator "1: Health-conscious" is assigned to that user.
[0038] Figure 5 shows that, as a result of the judgment rules, the person data of "User A" is assigned the behavioral factor indicators "1: Health-conscious", "0: Not a sweet tooth", and "1: Owns a car", the person data of "User B" is assigned the behavioral factor indicators "0: Not health-conscious", "0: Not a sweet tooth", and "1: Owns a car", and the person data of "User C" is assigned the behavioral factor indicators "1: Has a sweet tooth" and "0: Does not own a car". In addition, for the person data of "User C", "N / A: Undeterminable" is recorded as the result of the judgment rules for "Health-conscious" due to reasons such as insufficient purchase history being recorded to make a judgment according to the judgment rules.
[0039] Furthermore, the judgment rules may be defined for person data based on multiple perspectives. For example, they may be defined for "purchase history," "travel behavior," and "survey response history" based on multiple perspectives such as "purchase behavior," "travel behavior," and "survey response behavior." The judgment rules may be defined for each type of behavioral factor indicator, based on different perspectives for individual data. For example, the judgment rule for "health orientation" may be defined for "purchase history" based on "purchasing behavior," and the judgment rule for "car ownership" may be defined for "travel history" based on "travel behavior."
[0040] As shown in Figure 4, the control unit 30 functions as a data acquisition unit 300 and a machine learning unit 301 by executing a machine learning program 311 stored in the memory unit 41.
[0041] The data acquisition unit 300 receives requests from, for example, the machine learning unit 301 or the worker terminal device 5, and accesses the person database 20 via the communication unit 32 and the network 6. The data acquisition unit 300 acquires person data from the person database 20, recording the actions of multiple people, for each person, and provides it to the machine learning unit 301 or stores it in the storage unit 31.
[0042] The machine learning unit 301 generates training data 11 consisting of person data and behavioral factor indicators, and inputs multiple sets of this data into the learning model 12, thereby allowing the learning model 12 to learn the correlation between person data and behavioral factor indicators through machine learning. The machine learning unit 301 then stores the trained learning model 12 (specifically, the adjusted parameter set) in the storage unit 31.
[0043] Furthermore, when the machine learning unit 301 inputs the person data constituting the training data 11a into the input layer 120 of the learning model 12a, it may perform preprocessing to convert the person data into predetermined features for input into the learning model 12 using any conversion method such as one-hot encoding or label encoding. In this case, the preprocessing may be performed by appropriately combining various numerical calculations such as comparison, proportion, frequency, period, category, and behavioral feature indicators (product indicator, store indicator, website indicator, location indicator, regional indicator) with the purchase history, web browsing history, web search history, movement history, and survey response history included in the person data, thereby converting the person data into features.
[0044] In particular, if the person data is recorded in association with feature data (products, stores, websites, locations, regions) to which behavioral feature indicators (product indicators, store indicators, website indicators, location indicators, region indicators) are attached, as shown in Figure 3, the machine learning unit 301 may perform preprocessing to convert the person data into features based on the behavioral feature indicators attached to the feature data included in the person data, and then input multiple sets of preprocessed training data 11, which consists of the features converted from the person data in the preprocessing and the behavioral factor indicators included in the person data, into the learning model 12, thereby allowing the learning model 12 to learn the correlation between the person data and the behavioral factor indicators. As an example of the preprocessing in this case, the purchase amount of products with the behavioral factor indicator "health-conscious," the purchase amount of products with the behavioral factor indicator "sweet tooth," the purchase amount of products with the behavioral factor indicator "car owner," etc., are aggregated based on product indicators attached to products that the user has purchased in the past, which are recorded as part of the purchase history, and then each of the aggregated purchase amounts is normalized to a value in the range of 0 to 1 to be converted into features.
[0045] Figure 6 shows an example of training data 11 and a learning model 12. The training data 11 used for machine learning of the learning model 12 consists of person data and behavioral factor indicators. The training data 11 is used as training data, validation data, and test data in supervised learning. The behavioral factor indicators are used as correct labels in supervised learning.
[0046] The person data that makes up the training data 11 is obtained from the person database 20. The behavioral factor indicators that make up the training data 11 are the judgment rules registered in the judgment rule data 310. It may be assigned based on the judgment results, or it may be assigned by worker 10 through annotation work.
[0047] The learning model 12 employs, for example, a neural network structure and comprises an input layer 120, a hidden layer 121, and an output layer 122. Synapses (not shown) connect each neuron between each layer, and each synapse is associated with a weight.
[0048] The input layer 120 has a number of neurons corresponding to the person data (or features converted from the person data in preprocessing) as input data, and each value of the person data is input to each neuron. The output layer 122 has neurons corresponding to the behavioral factor index as output data, and the result of assigning the behavioral factor index to the person data (inference result) is output as output data (in this embodiment, a value in the range of 0 to 1). The learning model 12 is machine-learned by comparing the behavioral factor index (ground truth labels) that make up the training data 11 with the behavioral factor index (inference result) output from the output layer 122 and performing a process to adjust the parameter group such as the weight of each synapse (backpropagation, etc.).
[0049] In this embodiment, we will describe a case where the machine learning unit 301 performs machine learning on the learning model 12 for each type of behavioral factor indicator. Therefore, the memory unit 31 stores multiple learning models 12 corresponding to each type of behavioral factor indicator. The learning model 12 shown in Figure 6 outputs the behavioral factor indicator "health-conscious" for person data, and machine learning is performed using training data 11 consisting of person data and the behavioral factor indicator "health-conscious". Note that the learning model 12 may output multiple behavioral factor indicators for person data. For example, it may output three behavioral factor indicators: "health-conscious," "sweet tooth," and "car owner." In that case, machine learning should be performed using training data 11 consisting of person data and the behavioral factor indicators "health-conscious," "sweet tooth," and "car owner."
[0050] The number and types of learning models 12 that are machine-learned by the machine learning unit 301 and stored in the memory unit 31 may be changed as appropriate. For example, multiple learning models 12 with different conditions may be stored, such as machine learning methods, types of data included in the person data, types of data included in the behavioral factor indicators, and preprocessing methods for the person data. In that case, multiple types of training data 11, each having a data structure corresponding to the multiple learning models 12 with different conditions, should be used.
[0051] As shown in Figure 4, the machine learning unit 301 includes a first generation processing unit 301a, a first learning processing unit 301b, a second generation processing unit 301c, and a second learning processing unit 301d, each of which performs machine learning on the learning model 12.
[0052] Figure 7 is a functional diagram illustrating an example of the machine learning unit 301. Below, the details of the machine learning unit 301 will be explained with reference to Figure 7, the decision rule data 310 shown in Figure 5, and the training data 11 and training model 12 shown in Figure 6.
[0053] The first generation processing unit 301a determines, for each person, whether the person whose behavior is recorded in the person data acquired by the data acquisition unit 300 conforms to the behavior factor index, according to the determination rules registered in the determination rule data 310. Based on the determination result, the first generation processing unit 301a generates training data 11a by assigning the behavior factor index to the person data.
[0054] In the example in Figure 7, the first generation processing unit 301a determines whether each person (users A to C) conforms to the "health-conscious" criteria according to the judgment rules (see Figure 5) defined for the purchase record. Based on the judgment results, training data 11a is generated by assigning "1: Health-conscious" to the person data of "User A" and "0: Not health-conscious" to the person data of "User B" as behavioral factor indicators (provisional correct labels). However, for person data of "User C" that cannot be judged according to the judgment rule for "health-conscious" due to reasons such as insufficient record-keeping of purchase history, no behavioral factor indicator (provisional correct label) is assigned, and therefore training data 11a based on person data of "User C" is not generated.
[0055] The first learning processing unit 301b uses the training data 11a generated by the first generation processing unit 301a to train the learning model 12a on the correlation between person data and behavioral factor indicators. In this case, the person data constituting the training data 11a includes not only purchase history but also web browsing history, web search history, movement history, and survey response history, as shown in Figure 6. Alternatively, the first learning processing unit 301b may use the pre-processed training data 11a to train the learning model 12a on the correlation between person data (features) and behavioral factor indicators.
[0056] The second generation processing unit 301c receives correction instructions for the behavior factor indicators output from the learning model 12a by inputting person data (or features converted from the person data in preprocessing) into the learning model 12a, which has been trained by the first learning processing unit 301b to learn correlations. Specifically, as described above, the second generation processing unit 301c displays the behavior factor indicators output from the learning model 12a in association with the person data for each person, and receives correction instructions for the displayed behavior factor indicators when the worker 10 performs annotation work using the worker terminal device 5. Then, based on the received results, the second generation processing unit 301c generates training data 11b by adding behavior factor indicators to the person data.
[0057] In the example shown in Figure 7, the second generation processing unit 301c inputs the person data of each person (users A to C) into the learning model 12a, outputting behavioral factor indicators (inference results) of "0.9" for "user A", "0.6" for "user B", and "0.8" for "user C". When the second generation processing unit 301c displays an evaluation list on the worker terminal device 5 that associates the behavioral factor indicators (inference results) output from the learning model 12a with the person data, the worker 10, after visually reviewing the evaluation list, checks the person data of "user B". If the worker determines that "1: Health-oriented" is more appropriate than "0: Not health-oriented", which was assigned as a provisional correct label by the first generation processing unit 301a, the worker 10 issues a correction instruction to assign "1: Health-oriented" to "user B". Furthermore, if worker 10 checks the person data of "User C" and determines that "0: Not health-conscious" is more appropriate than "1: Health-conscious", it issues a correction instruction to assign "0: Not health-conscious" to "User C". The second generation processing unit 301c then receives these correction instructions and, based on the results of the reception, assigns "1: Health-conscious" to the person data of "User B" and "0: Not health-conscious" to the person data of "User C" as behavioral factor indicators (corrected correct labels), thereby generating the training data 11b.
[0058] Furthermore, when performing machine learning on the learning model 12a corresponding to the preprocessed training data 11a, the second generation processing unit 301c may accept modification instructions for at least one of the behavior factor indicators output from the learning model 12a and the behavior feature indicators attached to the feature data included in the person data, by inputting the features converted from the person data in preprocessing (preprocessed person data) into the learning model 12a, which has learned correlations by the first learning processing unit 301b. Specifically, as described above, the second generation processing unit 301c displays the behavior factor indicators output from the learning model 12a and the behavior feature indicators attached to the feature data included in the person data for each person, and the worker 10 performs annotation work using the worker terminal device 5, thereby allowing the displayed behavior factors The system accepts modification instructions for at least one of the causal index and the behavioral feature index. The second generation processing unit 301c then generates training data 11b by assigning behavioral factor indexes to the person data based on the received instructions, and may also modify the behavioral feature indexes assigned to the feature data.
[0059] The second learning processing unit 301d uses at least the training data 11b generated by the second generation processing unit 301c to train the learning model 12b on the correlation between person data and behavioral factor indicators. In this case, the second learning processing unit 301d may also use the training data 11a generated by the first generation processing unit 301a (in the example in Figure 7, training data 11a based on the person data of "User A") in addition to the training data 11b generated by the second generation processing unit 301c. The second learning processing unit 301d may also use the pre-processed training data 11b to train the learning model 12b on the correlation between person data (features) and behavioral factor indicators.
[0060] The machine learning unit 301 may repeatedly execute the following processes until the learning termination condition is met: when the second learning processing unit 301d has trained the learning model 12b on the correlation between person data and behavioral factor indicators, the second generation processing unit 301c inputs person data into the learning model 12b trained by the second learning processing unit 301d instead of the first learning processing unit 301b, receives a correction instruction for the behavioral factor indicators output from the learning model 12b, and generates training data 11b by assigning behavioral factor indicators to the person data based on the result of the received instruction; and the second learning processing unit 301d trains the learning model 12b using the training data 11b generated by at least the second generation processing unit 301c. Whether or not the learning termination condition is met is determined, for example, by whether or not the accuracy of the learning model 12 has reached a predetermined accuracy, or whether or not the number of training data 11 has reached a predetermined upper limit.
[0061] Furthermore, when performing machine learning on the learning model 12b corresponding to the preprocessed training data 11b, if the machine learning unit 301 fails to meet the predetermined learning termination conditions when the second learning processing unit 301d has trained the learning model 12b on the correlation between the person data (features) and the behavioral factor indicators, the second generation processing unit 301c inputs the features converted from the person data in preprocessing (preprocessed person data) into the learning model 12b trained by the second learning processing unit 301d instead of the first learning processing unit 301b, thereby generating the behavioral factor indicators output from the learning model 12b. The process of receiving a correction instruction for at least one of the feature data included in the person data and the behavioral feature index attached to it, and generating training data 11b by attaching the behavioral factor index to the person data based on the result of the instruction, and correcting the behavioral feature index attached to the feature data, and the process by which the second learning processing unit 301d trains the learning model 12b using the pre-processed training data 11b generated by at least the second generation processing unit 301c, may be repeated until the learning completion condition is met.
[0062] (Data processing device 4) Figure 8 is a block diagram showing an example of a data processing device 4. Figure 9 is a functional diagram showing an example of a data processing device 4. The data processing device 4 comprises a control unit 40 composed of a processor, a storage unit 41 composed of an HDD, SSD, memory, etc., a communication unit 42 which is a communication interface with the network 6, an input unit 43 composed of a keyboard, mouse, etc., and a display unit 44 composed of a display, etc. Note that the input unit 43 and the display unit 44 may be omitted.
[0063] The memory unit 41 stores the learning model 12 and the data processing program 410, as well as the operating system, other programs, various data, etc. 1 functions as a trained model memory unit that stores the trained model 12.
[0064] The control unit 40 functions as a data acquisition unit 400, an index assignment unit 401, and an output processing unit 402 by executing a data processing program 410 stored in the storage unit 41.
[0065] The data acquisition unit 400 receives requests, for example, from the indicator assignment unit 401 or the worker terminal device 5, and accesses the person database 20 via the communication unit 42 and the network 6. The data acquisition unit 400 acquires person data from the person database 20 and provides it to the indicator assignment unit 401 or stores it in the storage unit 41.
[0066] The index assignment unit 401 assigns behavioral factor indicators (inference results) output from the learning model 12 to the person data (which may be features converted from the person data in preprocessing) acquired by the data acquisition unit 400 to the learned learning model 12 stored in the memory unit 41. At that time, as post-processing for the behavioral factor indicators (inference results) output from the learning model 12 as output data (in this embodiment, values in the range of 0 to 1), the index assignment unit 401 compares the value of the behavioral factor indicator variable (in Figure 9, "0.9") with a predetermined threshold (for example, "0.5"). If the value is greater than or equal to the threshold, the behavioral factor indicator "1: Health-conscious" is assigned to the person data; if the value is less than the threshold, the behavioral factor indicator "0: Not health-conscious" is assigned to the person data. Furthermore, when the index assignment unit 401 inputs the person data to the input layer 120 of the learning model 12, it may perform the same preprocessing as the machine learning unit 301 so that the preprocessed person data, i.e., the features converted from the person data during preprocessing, are input to the learning model 12.
[0067] In this embodiment, since the learning model 12 is generated by the machine learning device 3 for each type of behavioral factor indicator, the memory unit 41 stores multiple learning models 12 corresponding to each type of behavioral factor indicator. The learning model 12 shown in Figure 9 outputs the behavioral factor indicator "health-oriented" for the person data. Therefore, the indicator assignment unit 401 assigns behavioral factor indicators other than "health-oriented" to the person data by inputting the person data into each of the multiple learning models 12 corresponding to each type of behavioral factor indicator, similar to Figure 9.
[0068] The output processing unit 402 performs output processing to output the behavior factor indicators assigned by the indicator assignment unit 401. The output processing unit 402 may, for example, provide the behavior factor indicators to the database device 2 to register them in the behavior factor indicator table of the person database 20, or provide them to the worker terminal device 5 to present them to the worker. In this case, the output processing unit 402 may output the behavior factor indicators before post-processing (values in the range of 0 to 1) in addition to or instead of the behavior factor indicators after post-processing by the indicator assignment unit 401 (0 or 1).
[0069] Furthermore, the number of learning models 12 stored in the memory unit 41 is not limited to one. For example, multiple learning models 12 with different conditions, such as machine learning methods, types of data included in person data, types of data included in behavioral factor indicators, and preprocessing methods for person data, may be stored and selectively used. In addition, the learning models 12 may be stored in the memory unit of an external computer (for example, a server-type computer or a cloud-type computer), in which case the data processing device 4 only needs to access that external computer.
[0070] (Machine learning methods) Figures 10 and 11 are flowcharts illustrating an example of a machine learning method using machine learning device 3. The following describes a case in which worker 10 performs a series of learning instruction operations, including annotation work, using worker terminal device 5, causing machine learning device 3 to refer to person database 20 and judgment rule data 310 and perform machine learning of learning model 12.
[0071] First, in step S100, when the operator terminal device 5 receives a learning instruction from the operator 10 that includes the learning conditions for performing machine learning on the learning model 12, it transmits the learning instruction to the machine learning device 3 in step S101. The learning conditions may include, for example, the conditions for acquiring person data when acquiring person data from the person database 20 to generate training data 11, and the conditions for ending the learning process.
[0072] Next, in step S110, when the machine learning device 3 receives the learning instruction for the learning model 12 transmitted in step S101, the data acquisition unit 300 accesses the person database 20 to acquire person data that satisfies the learning conditions of the learning model 12 included in the learning instruction, and acquires multiple person data corresponding to multiple people.
[0073] Next, in step S120, the first generation processing unit 301a determines for each person whether the behavior of each person conforms to the behavior factor index for the multiple person data acquired in step S110, according to the determination rules registered in the determination rule data 310. Then, in step S121, the first generation processing unit 301a generates multiple sets of training data 11a by assigning a behavior factor index to each person in the person data based on the determination result.
[0074] Next, in step S130, the first learning processing unit 301b uses the multiple sets of training data 11a generated in step S121 (which may also be pre-processed training data 11a) to train the learning model 12a on the correlation between person data and behavioral factor indicators.
[0075] Next, in step S140, the second generation processing unit 301c inputs the multiple person data acquired in step S110 into the learning model 12a that was trained in step S130, and associates the behavioral factor indicators output from the learning model 12a as inference results with the person data to create an evaluation list consisting of the person data and the behavioral factor indicators. Then, in step S141, the second generation processing unit 301c transmits the evaluation list to the worker terminal device 5.
[0076] Next, in step S142, when the worker terminal device 5 receives the evaluation list transmitted in step S141, it displays the evaluation list on the display screen of the worker terminal device 5. At that time, the worker terminal device 5 displays person data corresponding to the person displayed in the evaluation list on the display screen of the worker terminal device 5 as reference information for the worker 10 to perform the annotation work.
[0077] The display screen may show, for example, purchase history, web browsing history, web search history, movement history, and survey response history included in the person data. Furthermore, it may also display feature data associated with each history (in this embodiment, product, store, website, location, region) and behavioral feature indicators assigned to that feature data (in this embodiment, product indicator, store indicator, website indicator, location indicator, region indicator). For example, a list of products purchased by each person and the product attributes of each product may be displayed, and as part of the product attributes, the product indicators currently assigned to the product (e.g., "luxury-oriented") may be displayed. In addition, the display screen may display not only the raw data of the person data, but also information on which predetermined statistical processing has been performed on the person data, and the content of the display and the content of the statistical processing may be changed based on the operation of the operator 10. Furthermore, the evaluation list on the display screen may display not only the inference results by the learning model 12a in step S140, but also the judgment results by the judgment rules in step S120.
[0078] Then, in step S143, when the worker terminal device 5 receives a correction instruction for the behavioral factor indicator from worker 10 on its display screen, it transmits the correction instruction to the machine learning device 3 in step S144. In the example in Figure 11, it is shown that, as part of worker 10's annotation work, the worker terminal device 5 receives a correction instruction to assign "1: Health-conscious" to "User B" and a correction instruction to assign "0: Not health-conscious" to "User C," similar to Figure 7.
[0079] Next, in step S145, the second generation processing unit 301c receives the correction instruction sent in step S144. Upon receiving the correction instruction for the behavioral factor index output in step S140, it generates multiple sets of training data 11b by assigning the behavioral factor index to the person data based on the received result. The worker terminal device 5 may also receive correction instructions for behavioral feature indexes from the worker 10 on its display screen. In this case, the second generation processing unit 301c updates (corrects) the person database 20 by correcting the behavioral feature index in the feature data registered in the person database 20 based on the received result. Correction instructions for behavioral feature indexes (in this case, product indexes) may include, for example, adding the "healthy use" product index to a product that currently has the "luxury-oriented" behavioral feature index, or deleting the "luxury-oriented" product index.
[0080] Next, in step S150, the second learning processing unit 301d uses the training data 11b generated in step S145 (or the pre-processed training data 11b) to train the learning model 12b on the correlation between person data and behavioral factor indicators. At this time, the second learning processing unit 301d may also use the training data 11a generated in step S121, in addition to the training data 11b generated in step S145. If the person database 20 is updated based on the correction instructions for the behavioral feature indicators, the pre-processing of the training data 11b is performed with the updated person database 20 reflected.
[0081] Next, in step S160, the machine learning unit 301 determines whether a predetermined learning completion condition is met when the second learning processing unit 301d has trained the learning model 12b on the correlation between the person data and the behavioral factor indicators.
[0082] Then, if the machine learning unit 301 determines that the predetermined learning termination conditions are not met ("No" in step S160), it returns to step S140. At that time, in step S140, the second generation processing unit 301c inputs the multiple person data acquired in step S110 into the learning model 12b trained by the second learning processing unit 301d in step S150, instead of the learning model 12a trained by the first learning processing unit 301b in step S130. The machine learning unit 301 then repeatedly executes the subsequent processes (steps S140 to S150) until it determines in step S160 that the learning termination conditions are met.
[0083] On the other hand, if the machine learning unit 301 determines that the predetermined learning completion conditions are met (Yes in step S160), in step S170, the learning model 12b at that time is stored in the storage unit 31 as the learned learning model 12. This completes the series of processes shown in Figures 10 and 11. In the machine learning method, step S110 corresponds to the data acquisition process, steps S120 to S160 to the machine learning process, and step S170 to the learned model storage process.
[0084] In the above explanation, when the worker terminal device 5 receives a learning instruction from the worker 10, the series of processes shown in Figures 10 and 11 are executed, but the machine learning device 3 However, the machine learning unit 301 may execute a series of processes shown in Figures 10 and 11 when it determines that a predetermined execution condition (for example, when a certain amount of time has elapsed since the previous execution or when the person database 20 is updated) has been met. Also, although the above explanation states that if the machine learning unit 301 determines that the predetermined learning termination condition has not been met ("No" in step S160), it may return to step S120, or it may switch between returning to step S120 or step S140 based on the instructions of the operator 10 or the judgment of the machine learning unit 301. When returning to step S120, the operator 10 may, for example, change the judgment rule for the behavioral factor indicator, and the machine learning unit 301 may execute the processes from step S120 onward based on the changed judgment rule.
[0085] As described above, the machine learning device 3 and machine learning method according to this embodiment can provide a learning model 12 capable of estimating behavioral factor indicators that show the factors that motivate a person to take action, based on person data that records a person's actions from one or more perspectives. Therefore, by using this learning model 12, the data processing device 4 can appropriately assign (infer) behavioral factor indicators inherent in a person (user) without requiring the person (user) to input personal information.
[0086] Furthermore, the machine learning unit 301 performs machine learning using the training data 11a generated based on the judgment results of the judgment rules, and then performs machine learning in stages using the training data 11b generated based on the results of receiving correction instructions for the behavioral factor indicators, thereby training the learning model 12 on the correlation between person data and behavioral factor indicators. As a result, the machine learning of the learning model 12 is performed in stages through the automatic generation of training data 11a by the judgment rules and the correction of training data 11b by the worker 10's annotation work, thus reducing the workload of the worker 10's annotation work while generating a highly accurate learning model 12.
[0087] Furthermore, the machine learning unit 301 updates the behavioral feature indicators attached to the feature data in the person database 20 based on the results of receiving correction instructions for the behavioral feature indicators, thereby performing machine learning with the updated person database 20 reflected. As a result, the accuracy of the behavioral feature indicators attached to the feature data included in the person data is improved through the correction of behavioral feature indicators by the worker 10's annotation work, making it possible to generate a more accurate learning model 12.
[0088] (Data processing method) Figure 12 is a flowchart showing an example of a data processing method by the data processing device 4. Below, we will describe a case in which an operator 10 performs a tag assignment instruction operation using the operator terminal device 5, and the data processing device 4 refers to the person database 20 and the trained learning model 12 to assign behavioral factor indicators to the person data.
[0089] First, in step S200, when the worker terminal device 5 receives an assignment instruction from the worker 10, which includes the assignment conditions for assigning behavioral factor indicators, it transmits the assignment instruction to the machine learning device 3 in step S201. The assignment conditions include, for example, the conditions for acquiring person data when acquiring person data to which behavioral factor indicators will be assigned from the person database 20.
[0090] Next, in step S210, when the data processing device 4 receives the instruction to assign behavioral factor indicators transmitted in step S201, the data acquisition unit 400 accesses the person database 20 to acquire person data that satisfies the assignment conditions for the behavioral factor indicators included in the assignment instruction, and acquires, for example, multiple person data corresponding to multiple people.
[0091] Next, in step S220, the index assignment unit 401 inputs the multiple person data acquired in step S110 into the learning model 12 and assigns the behavioral factor indicators output from the learning model 12 to the person data as inference results. In the example in Figure 12, the index assignment unit 401 performs post-processing based on a comparison of the variable values of the behavioral factor indicators with predetermined thresholds, and assigns "1: Health-oriented" to "User P" and "User R," while assigning "0: Not health-oriented" to "User Q." The assignment list is shown in this case.
[0092] Next, in step S230, the output processing unit 402 provides the behavioral factor indicators (assigned list in Figure 12) that were assigned to the person data in step S220 to, for example, the database device 2, so that they are registered in the behavioral factor indicator table of the person database 20. This completes the series of processes shown in Figure 12. In the data processing method, step S210 corresponds to the data acquisition step, step S220 to the indicator assignment step, and step S230 to the output processing step.
[0093] In the above explanation, it was described that the series of processes shown in Figure 12 are executed when the worker terminal device 5 receives an instruction from the worker 10. However, the data processing device 4 may also be configured to execute the series of processes shown in Figure 12 when it determines that a predetermined execution condition (for example, when a certain amount of time has elapsed since the previous execution or when the person database 20 is updated) is met.
[0094] As described above, the data processing device 4 and data processing method according to this embodiment make it possible to appropriately assign (infer) behavioral factor indicators inherent to a person (user) without requiring the person to input personal information. The behavioral factor indicators assigned to each person can then be processed by predetermined statistical processing and visualization processing, etc., and used in various services (product planning, product development, production management, procurement and ordering management, sales strategy formulation, marketing, advertising, etc.).
[0095] (Other embodiments) The present invention is not limited to the embodiments described above, and can be implemented with various modifications without departing from the spirit of the invention. All such modifications are included in the technical concept of the present invention.
[0096] In the above embodiment, the database device 2, the machine learning device 3, and the data processing device 4 were described as being composed of separate devices, but these three devices may be composed of a single device, or any two of these three devices may be composed of a single device. Furthermore, at least one of the machine learning device 3 and the data processing device 4 may be incorporated into the worker terminal device 5.
[0097] In the above embodiment, a case in which a neural network is used as the learning model for realizing machine learning by the machine learning unit 301 was described, but other machine learning models may also be used. Other machine learning models include, for example, tree models such as decision trees and regression trees, ensemble learning models such as bagging and boosting, recurrent neural networks, convolutional neural networks, and neural network models such as LSTM (deep (including learning), hierarchical clustering, non-hierarchical clustering, k-nearest neighbors, k-squared Examples include clustering models such as the equalization method, multivariate analysis models such as principal component analysis, factor analysis, and logistic regression, and support vector machines.
[0098] (Inference device, inference method, or inference program) The present invention can be provided not only in the form of the data processing device 4 (data processing method or data processing program) according to the above embodiment, but also in the form of an inference device (inference method or inference program). In that case, the inference device (inference method or inference program) may include a memory and a processor, the processor of which may execute a series of processes. The series of processes includes a data acquisition process (data acquisition step) for acquiring person data, and an inference process (inference step) for inferring behavioral factor indicators from the person data acquired in the data acquisition process.
[0099] By providing the inference device (inference method or inference program) in the form of an inference device, it becomes easier to apply to various devices compared to implementing the data processing device 4. It will be obvious to those skilled in the art that when the inference device (inference method or inference program) infers behavioral factor indicators, it may apply the inference method performed by the indicator assignment unit using the trained learning model generated by the machine learning device 3 and machine learning method according to the above embodiment. [Explanation of symbols]
[0100] 1...Person information management system, 2...Database device, 3...Machine learning device, 4...Data processing device, 5...Worker terminal device, 6...Network, 10...Worker, 11, 11a, 11b...Training data, 12, 12a, 12b...Learning models, 20...People database, 30...Control unit, 31...Storage unit, 32...Communication unit, 33...Input unit, 34...Display unit, 40...Control unit, 41...Storage unit, 42...Communication unit, 43...Input unit, 44...Display unit, 120...Input layer, 121...Hidden layer, 122...Output layer, 300...Data acquisition unit, 301...Machine learning unit, 301a...First generation processing unit, 301b...First learning processing unit, 301c...Second generation processing unit, 301d...Second learning processing unit, 310... Decision rule data, 311... Machine learning program, 400...Data acquisition unit, 401...Index assignment unit, 402...Output processing unit, 410...Data processing program
Claims
1. a machine learning unit that inputs a plurality of sets of learning data, each set being composed of person data that records a person's behavior based on one or more viewpoints and a behavioral factor index that indicates a factor that motivates the person to perform the behavior, into a learning model, and causes the learning model to learn a correlation between the person data and the behavioral factor index by machine learning; A trained model storage unit that stores the trained model in which the correlation is trained by the machine learning unit, The machine learning unit is a first generation processing unit that generates the learning data by determining, for each of the persons, whether or not the person matches the behavioral factor index in accordance with a determination rule defined for the person data based on at least a part of the viewpoint, and assigning the behavioral factor index to the person data based on the determination result; a first learning processing unit that causes the learning model to learn the correlation by using the learning data generated by the first generation processing unit; a second generation processing unit that receives a correction instruction for the behavior factor index output from the learning model by inputting the person data into the learning model in which the correlation has been learned by the first learning processing unit, and generates the learning data by assigning the behavior factor index to the person data based on the reception result; A second learning processing unit that causes the learning model to learn the correlation by using the learning data generated by at least the second generation processing unit. Machine learning device.
2. The second generation processing unit, inputting the person data into the learning model in which the correlation has been learned by the first learning processing unit, thereby causing the behavioral factor indexes output from the learning model to be associated with the person data and displayed for each person, and receiving the modification instruction for the displayed behavioral factor indexes; The machine learning device according to claim 1 .
3. The second generation processing unit, The person data is input to the learning model in which the correlation is learned by the first learning processing unit. and displaying the behavioral factor indexes output from the learning model for each person by inputting the data, in association with information obtained by performing a predetermined statistical process on the person data, and receiving the instruction to modify the displayed behavioral factor indexes. The machine learning device according to claim 1 .
4. The second generation processing unit, the behavior factor index assigned to the person data by the first generation processing unit based on the determination result, and the behavior factor index output from the learning model in which the correlation has been learned by the first learning processing unit by inputting the person data into the learning model, are displayed for each person in association with the person data, and the modification instruction is received for the displayed behavior factor index; The machine learning device according to claim 1 .
5. The machine learning unit is When the second learning processing unit has learned the correlation in the learning model, if a predetermined learning end condition is not satisfied, the second generation processing unit inputs the person data into the learning model in which the correlation has been learned by the second learning processing unit instead of the first learning processing unit, accepts the correction instruction for the behavior factor index output from the learning model, and generates the learning data by assigning the behavior factor index to the person data based on the acceptance result; and the second learning processing unit repeatedly executes a process of having the learning model learn the correlation using at least the learning data generated by the second generation processing unit until the learning end condition is satisfied. The machine learning device according to claim 1 .
6. a machine learning unit that inputs a plurality of sets of learning data, each set being composed of person data that records a person's behavior based on one or more viewpoints and a behavioral factor index that indicates a factor that motivates the person to perform the behavior, into a learning model, and causes the learning model to learn a correlation between the person data and the behavioral factor index by machine learning; A trained model storage unit that stores the trained model in which the correlation is trained by the machine learning unit, The person data is a behavioral characteristic index corresponding to the behavior factor index is recorded in association with the assigned characteristic data, The machine learning unit is a pre-processing is performed to convert the personal data into features for input to the learning model based on the behavioral feature index assigned to the feature data included in the personal data, and a plurality of sets of pre-processed learning data, which are composed of the features converted from the personal data in the pre-processing and the behavioral factor index included in the personal data, are input to the learning model, thereby causing the learning model to learn the correlation; The machine learning unit is a first generation processing unit that generates the learning data by determining, for each of the persons, whether or not the person matches the behavioral factor index in accordance with a determination rule defined for the person data based on at least a part of the viewpoint, and assigning the behavioral factor index to the person data based on the determination result; a first learning processing unit that causes the learning model to learn the correlation by using the learning data generated by the first generation processing unit and the learning data after the preprocessing; The feature amount converted from the person data in the preprocessing is input to the learning model in which the correlation is learned by the first learning processing unit, and the behavior factor index output from the learning model and the preprocessing data assigned to the feature data included in the person data are output. a second generation processing unit that receives a correction instruction for at least one of the behavioral feature index and the behavior factor index, and generates the learning data by assigning the behavior factor index to the person data based on the reception result, and corrects the behavioral feature index assigned to the feature data; a second learning processing unit that causes the learning model to learn the correlation by using the learning data generated by at least the second generation processing unit, the learning data being the preprocessed learning data; Machine learning device.
7. The second generation processing unit, inputting the feature amount converted from the person data in the preprocessing into the learning model in which the correlation has been learned by the first learning processing unit, the behavior factor index output from the learning model and the behavior feature index assigned to the feature data included in the person data are displayed for each person in association with the person data, and the modification instruction is received for at least one of the displayed behavior factor index and the behavior feature index. The machine learning device according to claim 6.
8. The second generation processing unit, the feature amount converted from the person data in the preprocessing is input to the learning model in which the correlation has been learned by the first learning processing unit, and the behavior factor index output from the learning model and the behavior feature index assigned to the feature data included in the person data are displayed for each person in association with information obtained by performing a predetermined statistical process on the person data, and the modification instruction is received for at least one of the displayed behavior factor index and the behavior feature index. The machine learning device according to claim 6.
9. The second generation processing unit, the behavior factor index assigned to the person data by the first generation processing unit based on the determination result, and the feature amount converted from the person data in the preprocessing are input to the learning model in which the correlation has been learned by the first learning processing unit, and the behavior factor index output from the learning model and the behavior feature index assigned to the feature data included in the person data are associated with the person data and displayed for each person, and the modification instruction is received for at least one of the behavior factor index and the behavior feature index. The machine learning device according to claim 6.
10. The machine learning unit is When the second learning processing unit has learned the correlation in the learning model, if a predetermined learning end condition is not satisfied, the second generation processing unit receives the modification instruction for at least one of the behavior factor index output from the learning model, in which the correlation has been learned by the second learning processing unit instead of the first learning processing unit, by inputting the feature amount converted from the person data in the preprocessing, and the behavior feature index assigned to the feature data included in the person data, and generates the learning data by assigning the behavior factor index to the person data based on the reception result, and modifies the behavior feature index assigned to the feature data; and the second learning processing unit repeatedly executes a process of having the learning model learn the correlation using at least the learning data generated by the second generation processing unit and the learning data after the preprocessing, until the learning end condition is satisfied. The machine learning device according to any one of claims 6 to 9.
11. a machine learning process for inputting a plurality of sets of learning data, each set being composed of person data that records a person's behavior based on one or more viewpoints and a behavioral factor index that indicates a factor that motivates the person to perform the behavior, into a learning model, thereby causing the learning model to learn a correlation between the person data and the behavioral factor index by machine learning; A trained model storage step of storing the trained model in a trained model storage unit, the trained model having the correlation trained by the machine learning step; The machine learning process includes: a first generation process step of judging for each of the persons whether or not the person matches the behavioral factor index in accordance with a judgment rule defined for the person data based on at least a part of the viewpoint, and generating the learning data by assigning the behavioral factor index to the person data based on the judgment result; a first learning process step of causing the learning model to learn the correlation by using the learning data generated by the first generation process step; a second generation processing step of receiving a correction instruction for the behavior factor index output from the learning model by inputting the person data into the learning model that has learned the correlation by the first learning processing step, and generating the learning data by assigning the behavior factor index to the person data based on the reception result; and a second learning process step of causing the learning model to learn the correlation by using the learning data generated at least by the second generation process step. Machine learning methods.
12. a machine learning process for inputting a plurality of sets of learning data, each set being composed of person data that records a person's behavior based on one or more viewpoints and a behavioral factor index that indicates a factor that motivates the person to perform the behavior, into a learning model, thereby causing the learning model to learn a correlation between the person data and the behavioral factor index by machine learning; A trained model storage step of storing the trained model in a trained model storage unit, the trained model having the correlation trained by the machine learning step; The person data is a behavioral characteristic index corresponding to the behavior factor index is recorded in association with the assigned characteristic data, The machine learning process includes: a pre-processing is performed to convert the personal data into features for input to the learning model based on the behavioral feature index assigned to the feature data included in the personal data, and a plurality of sets of pre-processed learning data, which are composed of the features converted from the personal data in the pre-processing and the behavioral factor index included in the personal data, are input to the learning model, thereby causing the learning model to learn the correlation; The machine learning process includes: a first generation process step of judging for each of the persons whether or not the person matches the behavioral factor index in accordance with a judgment rule defined for the person data based on at least a part of the viewpoint, and generating the learning data by assigning the behavioral factor index to the person data based on the judgment result; a first learning process step of causing the learning model to learn the correlation by using the learning data generated by the first generation process step and the learning data after the preprocessing; The feature amount converted from the person data in the preprocessing is input to the learning model that has learned the correlation in the first learning processing step, and a correction instruction is received for at least one of the behavior factor index output from the learning model and the behavior feature index assigned to the feature data included in the person data, and the learning data is generated by assigning the behavior factor index to the person data based on the reception result. and a second generation process step of correcting the behavioral feature index assigned to the feature data. and a second learning process step of causing the learning model to learn the correlation by using the learning data generated by at least the second generation process step, the learning data being the preprocessed learning data. Machine learning methods.
