Apparatus and method for proposing an action

Abstract

To appropriately suggest actions to participants of an interaction group.SOLUTION: An apparatus extracts feature quantities of a first participant from interaction data including text messages posted by multiple participants of a first interaction group. The apparatus calculates a predetermined score of the first participant based on at least a part of the feature quantities of the first participant. The apparatus estimates a variation value of the predetermined score due to each improvement action candidate for improving the predetermined score based on the predetermined score. The improvement action candidate includes participation in a different interaction group from the interaction group. The apparatus determines a suggested improvement action for the first participant based on the variation value of the predetermined score.SELECTED DRAWING: Figure 7

Description

Technical Field

[0001] The present invention relates to a technique for proposing actions to participants in an AC group.

Background Art

[0002] From the perspective of ESG (Environment, Social and Governance) investment, companies that engage in human capital investment and healthy management are attracting attention. The issue is how to reduce the mental leave and presenteeism loss costs that directly affect corporate costs. On the other hand, in the case of telecommuting, it is becoming even more difficult to grasp the stress and productivity of each employee.

[0003] As a related technique of the present disclosure, there is Japanese Patent Application Laid-Open No. 2020-57067 (Patent Document 1). Patent Document 1 discloses a method of analyzing a user's communication without using a wearable device. Analyze the chat and calculate scores for group activity, engagement, and condition.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] Patent Document 1 evaluates group activity, engagement, and condition from chat data. However, with only these evaluations, it is not possible to directly improve the stress and productivity of the target person. Therefore, a technique that can propose appropriate actions to the participants in the group is desired.

Means for Solving the Problems

[0006] To solve the above problems, a method for proposing actions to participants in an exchange group according to one aspect of the present invention includes: an apparatus extracting features of a first participant from exchange data including text messages posted by multiple participants of a first exchange group; the apparatus calculating a predetermined score for the first participant based on at least a portion of the features of the first participant; the apparatus estimating the change in the predetermined score for each candidate improvement action that improves the predetermined score based on the predetermined score; and the candidate improvement action is the 1 The device determines a suggested improvement action for the first participant based on the change in the predetermined score, including participation in a different exchange group than the one already in place.

[0007] Further features relating to the present invention will become apparent from the description herein and the accompanying drawings. Furthermore, aspects of the present invention are achieved and realized by elements and various combinations of elements and by the modes of the claims described herein in detail and the accompanying claims.

[0008] It should be understood that the descriptions herein are merely typical examples and do not limit the claims or applications of the present invention in any way. [Effects of the Invention]

[0009] According to one aspect of the present invention, appropriate actions can be suggested to participants in an exchange group. [Brief explanation of the drawing]

[0010] [Figure 1] This block diagram shows an example of the functional configuration of a system that supports stress reduction and productivity improvement through chat analysis. [Figure 2] This diagram shows an example of the hardware configuration of a chat analysis device. [Figure 3] This shows an example of the training data structure for the employee classification estimation model and the score estimation model. [Figure 4] This shows an example of the structure of training data for a score improvement behavior estimation model. [Figure 5] Shows a configuration example of an AC group DB. [Figure 6] Shows a configuration example of an AC data DB. [Figure 7] Shows a flowchart of a processing example of a chat analysis device. [Figure 8] Shows an example of input data to a score improvement action estimation model. [Figure 9] Shows a configuration example of an improvement score table. [Figure 10A] Shows an example of a screen displayed on an employee terminal. [Figure 10B] Shows an example of a screen displayed on an employee terminal. [Figure 10C] Shows an example of a screen displayed on an employee terminal. [Figure 10D] Shows an example of a screen displayed on an employee terminal.

Mode for Carrying Out the Invention

[0011] Hereinafter, embodiments of this specification will be described with reference to the accompanying drawings. In the accompanying drawings, functionally identical elements may sometimes be denoted by the same number. Note that the accompanying drawings show specific embodiments and implementation examples in accordance with the principles of the present invention, but these are for the purpose of understanding the present invention and are not used to limit the interpretation of the present invention in any way.

[0012] In this embodiment, although the description is made in sufficient detail for those skilled in the art to implement the present invention, other implementations and forms are also possible, and it is necessary to understand that changes in configuration and structure and replacement of various elements are possible without departing from the scope and spirit of the technical idea of the present invention. Therefore, the following description should not be interpreted as being limited thereto.

[0013] Furthermore, the embodiments of this specification may be implemented by software running on a general-purpose computer as described later, or may be implemented by dedicated hardware or a combination of software and hardware.

[0014] When explaining each process in the embodiments of this specification with "each processing unit as a program" as the subject (the operating entity), since the program performs the processes defined by being executed by a processor (such as a CPU) while using a memory and a communication port (communication control device), it may be described with the processor as the subject. <System Configuration>

[0015] FIG. 1 is a block diagram showing a functional configuration example of a system that supports stress improvement and productivity improvement by chat analysis according to an embodiment of this specification. According to this embodiment, it is possible to effectively improve the stress and productivity of the target person, and as a result, to activate the organization and reduce the mental illness rate. In the chat system, a plurality of participants can post (send) messages in real time and read (receive) the posted messages. Note that the method described below can also be applied to a system or service different from chat, such as a bulletin board system, in which a plurality of participants exchange text messages. Further, it can be used to improve the state of participants different from stress or productivity.

[0016] The system shown in FIG. 1 is a computer system. The computer system includes a chat analysis device 1 and an employee terminal 3, which communicate via a network 5. FIG. 1 shows one employee terminal 3 as an example, but the employee terminals of a plurality of employees participating in the chat are connected to the network 5.

[0017] The chat analysis device 1 includes an employee classification estimation unit 10, a stress / productivity score estimation unit 11, an estimation result integration processing unit 12, a proposed action creation unit 13, a proposed action interest / effect determination unit 14, a model training unit 15, an employee classification estimation model 16, a score estimation model 17, a score improvement action estimation model 18, and a communication data collection unit 19. The chat analysis device 1 further stores a training data DB (database) 21, a communication group DB 23, an improvement action DB 25, and a communication data DB 27.

[0018] The employee classification estimation unit 10 uses the employee classification estimation model 16 to estimate the employee's personality type based on the chat content. The stress and productivity score estimation unit 11 uses the score estimation model 17 to estimate the employee's stress score and productivity score based on the chat content. Other types of scores may be used to improve other aspects of the employee's condition.

[0019] The estimation result integration processing unit 12 generates data to be input to the score improvement behavior estimation model 18. The proposed behavior creation unit 13 determines the behavior to propose to the employee based on the stress score and productivity score. The proposed behavior interest and effectiveness determination unit 14 determines the employee's interest in and effectiveness of the proposed behavior based on their response to that behavior.

[0020] The model training unit 15 trains the employee classification estimation model 16, the score improvement behavior estimation model 18, and the score estimation model 17. The employee classification estimation model 16 estimates the personality type of an employee based on the employee's chat data. The score estimation model 17 calculates the employee's productivity score and stress score based on the employee's chat data and employee classification.

[0021] The score improvement behavior estimation model 18 estimates suggested behaviors that may contribute to improving an employee's score, based on their productivity score and stress score. The interaction data collection unit 19 collects information on interaction groups and chat data from each interaction group from employee terminals 3.

[0022] Employee terminal 3 includes a chat input / transmission unit 31 and a chat display unit 32. The chat input / transmission unit 31 receives chat input from the employee and transmits it to a server (not shown) that provides the chat service. The chat display unit 32 displays the chat screen between the employee and other employees, as well as a screen showing improvement actions suggested by the chat analysis device 1. Improvement actions may be displayed on the chat screen with other employees, the chat screen with the chat analysis device 1, or on the screen of another application.

[0023] The functional units 10 to 18 in the chat analysis device 1 and the functional units 31 and 32 in the employee terminal 3 may be implemented by a processor that executes programs, or by hardware modules implemented to match each functional unit.

[0024] Figure 2 shows an example of the hardware configuration of the chat analysis device 1 according to the embodiment of this specification. The hardware configuration example of the chat analysis device 1 will be described below, but employee terminals may have a similar configuration.

[0025] The chat analysis device 1 includes a CPU (processor) 201 that executes various programs, a memory (main memory) 202 that stores various programs, and an auxiliary storage device 203 that stores various data. The CPU 201 may include one or more cores, and the memory 202 is, for example, RAM that includes volatile storage area. The auxiliary storage device 203 is, for example, an HDD (hard disk drive) or flash memory, and can be provided with non-volatile storage area.

[0026] The chat analysis device 1 further includes an output device 204 for presenting information to the user, an input device 205 for inputting user instructions, images, etc., and a communication device 206 for communicating with other devices. These are interconnected by a bus 207.

[0027] The functional unit of the chat analysis device 1 shown in Figure 1 can be implemented, for example, by the CPU 201 operating according to a program. The CPU 201 reads and executes various programs from memory 202 as needed. Memory 202 stores the programs. In Figure 2, the same names and symbols as those for the functional unit shown in Figure 1 are assigned to the programs.

[0028] Specifically, memory 202 stores the programs for the employee classification estimation unit 10, the stress and productivity score estimation unit 11, the estimation result integration processing unit 12, the proposed action creation unit 13, the proposed action interest and effectiveness determination unit 14, the model training unit 15, the employee classification estimation model 16, the score estimation model 17, the score improvement action estimation model 18, and the interaction data collection unit 19. Each program is loaded into memory 202 from, for example, the auxiliary storage device 203 and executed by the CPU 201. At least a portion of the functional unit may be composed of logic circuits.

[0029] The auxiliary storage device 203 stores data that is referenced or managed by various programs. In Figure 2, the auxiliary storage device 203 stores the training data DB 21, the interaction group DB 23, the improvement behavior DB 25, and the interaction data DB 27.

[0030] The output device 204 consists of devices such as a display, printer, and speaker. The input device 205 consists of devices such as a keyboard, mouse, and microphone. The output device 204 displays the input results from the user as well as the processing results from the chat analysis device 1. User instructions are input to the chat analysis device 1 via the input device 205.

[0031] The communication device 206 receives data transmitted from other devices connected via a network, including, for example, employee terminal 3, and transmits the processing results from the chat analysis device 1 to other devices. Some devices may be omitted. <Training the model>

[0032] Figure 3 shows an example of the configuration of training data 310 for employee classification estimation model 16 and score estimation model 17. The training data 310 is included in training data DB 21. Employee classification estimation model 16 estimates the personality type of an employee from information obtained from chats in which the employee participates. Score estimation model 17 calculates the employee's stress score and productivity score from information obtained from chats in which the employee participates.

[0033] In the configuration example shown in Figure 3, the training data 310 represents the explanatory variables input to the employee classification estimation model 16 and the score estimation model 17, respectively, and the target variable output to either the employee classification estimation model 16 or the score estimation model 17. The explanatory variables are the features input to the model.

[0034] Specifically, the training data 310 includes the employee ID column 311 and the period column 312. The explanatory variable column includes the number of posts column 313, the sentiment analysis column for posts 314, the number of predictive words column 315, the average time of posts column 316, and the number of responses to posts column 317. The dependent variable column includes the employee classification column 318, the stress score column 319, and the productivity score column 320.

[0035] Each record in training data 310 represents information about a single employee in a chat within a specific interaction group over a specific period. The interaction group is a chat group. Multiple participants can optionally post messages in the chat space of the interaction group, and all participants can see the posted messages.

[0036] When an explanatory variable for one record is input to the employee classification estimation model 16, the employee classification estimation model 16 outputs the estimated employee classification. When an explanatory variable for one record is input to the score estimation model 17, the score estimation model 17 outputs estimated values ​​for stress score and productivity score.

[0037] The Employee ID column 311 shows the employee's ID. The Period column 312 shows the duration of the chat for each record. The Number of Posts column 313 shows the number of posts made by the employee indicated in the Employee ID column 311 during the period indicated in the Period column 312. The Post Sentiment Analysis column 314 shows the employee's sentiment derived from the analysis of the post content. In the example in Figure 3, sentiment is represented by "+" for positive or "-" for negative. It should be noted that estimating sentiment from text is possible, for example, using a machine learning model based on deep learning. Sentiment recognition technology is widely known, so details are omitted.

[0038] The "Number of Warning Words" column 315 shows the number of warning words appearing in the post. Warning words are, for example, warning words related to mental illness, and words identical or similar to pre-set words are extracted. The "Average Post Time" column 316 shows the average posting time of the message. The "Number of Post Reactions" column 317 shows the number of reactions to other people's posts. For example, "likes" are considered reactions.

[0039] The number of posts, average posting time, and number of post responses indicate the posting style of employees. The post sentiment analysis results and the number of predictive words are values ​​based on the content of the employee's posts. This information allows for a more accurate estimation of the employee's state. While it is possible to estimate an employee's state using only one type of information, using both types of information enables a more accurate estimation.

[0040] The employee classification column 318 indicates the employee's personality type. The stress score column 319 shows the employee's stress score calculated from the explanatory variables in the corresponding record. A lower stress score indicates less stress. The productivity score column 320 shows the employee's productivity score calculated from the explanatory variables in the corresponding record. A higher productivity score indicates higher productivity.

[0041] The model training unit 15 uses the training data 310 to train the employee classification estimation model 16 and the score estimation model 17. For example, these models 16 and 18 are composed of deep neural networks, and the model training unit 15 optimizes the parameters of models 16 and 18 using backpropagation.

[0042] Note that the explanatory variables shown in Figure 3 are examples, and some of them may be omitted or other explanatory variables may be added. For example, the amount of change from the average posting time in the previous aggregation may be added, or replaced with the average posting time. Alternatively, the amount of change in the number of post responses may be added, or replaced with the number of post responses. The employee classification estimation model 16 and the score estimation model 17 may be rule-based models different from machine learning models.

[0043] As described above, the inputs to the classification model and the scoring model can include variables based on the content of employee chat posts, posting frequency, length, sentiment analysis results, chat responses to suggested actions, and their fluctuations. These explanatory variables represent changes in employee mood and their willingness to post in chat, and can contribute to a more accurate estimation of stress scores and productivity scores. Note that at least some of the explanatory variables in the inputs to the two models may be omitted, and other types of variables may be added.

[0044] Figure 4 shows an example of the structure of the training data 350 for the score improvement behavior estimation model 18. The training data 350 is included in the training data DB 21. The score improvement behavior estimation model 18 estimates behaviors that can improve an employee's stress score and productivity score based on the employee classification and the employee's stress score and productivity score. By referring to both the classification and the score, a more accurate estimation is possible. Some of these may be omitted.

[0045] In the configuration example shown in Figure 4, the training data 350 represents the explanatory variables input to the score improvement behavior estimation model 18 and the target variable output to the score improvement behavior estimation model 18. Specifically, the training data 350 includes the employee ID column 351. The explanatory variable column includes the classification column 352, the productivity score column at the time of behavior suggestion 353, the stress score column at the time of behavior suggestion 354, the suggested behavior column 355, the group size column 356, the percentage of chat posters column 357, and the average number of chats per day column 358. The target variable column includes the productivity score fluctuation column 359 and the stress score fluctuation column 360.

[0046] Each record in the training data 350 represents information about a single employee in a chat within a specific interaction group over a specific period. When the explanatory variables of a single record are input into the score improvement behavior estimation model 18, the score improvement behavior estimation model 18 outputs estimates of score stress score variability and productivity score variability.

[0047] The Employee ID column 351 shows the employee's ID. The Classification column 352 shows the employee's personality type. The Productivity Score column 353 shows the employee's productivity score when the Score Improvement Behavior Estimation Model 18 estimates an improvement behavior. The Stress Score column 354 shows the employee's stress score when the Score Improvement Behavior Estimation Model 18 estimates an improvement behavior.

[0048] The suggested action column 355 shows the improvement actions suggested by the score improvement action estimation model 18. In this example, it shows the category of the suggested improvement action. For example, "Walking" means suggesting walking, and "Sleeping" means suggesting sleeping. "Information exchange group suggestion" means suggesting participation in other exchange groups.

[0049] The score improvement behavior estimation model 18 is input with the category of the proposed improvement behavior or the identifier of the interaction group loop to which participation is proposed. Including participation in other interaction groups as a candidate for proposed behavior can increase the likelihood of improving the employee's score.

[0050] The "Group Size" column 356 indicates the number of members in the group. The "Chat Posting Percentage" column 357 indicates the percentage of participants who posted a message within a specified period. Here, any participant who made at least one post is counted as a poster. Alternatively, only participants who posted more than a specified number of posts may be counted. The "Average Chat Frequency / Day" column 358 indicates the average number of posts per day in the group within a specified period. The unit of time for the average frequency may be a length other than one day. The specified period for calculating the values ​​in different columns may be the same or different.

[0051] The Productivity Score Change column 359 shows the change in productivity score from the productivity score at the time of the action suggestion, calculated from the explanatory variables of the corresponding record. The Stress Score Change column 360 shows the change in stress score from the stress score at the time of the action suggestion, calculated from the explanatory variables of the corresponding record. A negative stress score change indicates that the employee's mental state has improved.

[0052] By inputting data from different proposed behaviors into the score improvement behavior estimation model 18, score fluctuation values ​​for those different proposed behaviors can be obtained. By selecting behaviors that favorably change the score and proposing them to employees, improvements in employee stress and productivity can be expected.

[0053] The model training unit 15 trains the score improvement behavior estimation model 18 using the training data 350. For example, the score improvement behavior estimation model 18 is composed of a deep neural network, and the model training unit 15 optimizes the parameters of the score improvement behavior estimation model 18 using backpropagation.

[0054] Note that the explanatory variables shown in Figure 4 are examples, and some of them may be omitted or other explanatory variables may be added. The score improvement behavior estimation model 18 may also be a rule-based model different from the machine learning model.

[0055] The input to the score improvement behavior estimation model18 includes information on interaction groups in addition to the employee's score. This makes it possible to more accurately estimate the appropriate interaction groups for each employee. <Action suggestion>

[0056] The following describes an example of the action suggestion process performed by the chat analysis device 1 for employees. The chat analysis device 1 stores an employee classification estimation model 16, a score improvement action estimation model 18, and a score estimation model 17, which were trained by the model training unit 15 using the training data 310 and 350 from the training data DB 21. The chat analysis device 1 analyzes the employee's chat and suggests actions to improve the employee's stress score and / or productivity score.

[0057] The interaction data collection unit 19 of the chat analysis device 1 collects chat data from each employee's interaction group in order to provide action suggestions. The interaction data collection unit 19 analyzes the collected data and updates the interaction group DB 23 and the interaction data DB 27 based on the analysis results.

[0058] Figure 5 shows an example of the configuration of the Communication Group DB23. The Communication Group DB23 manages information about communication groups. Multiple employees belong to a communication group and participate in the chat of that communication group. Employees chat with other employees in the communication group to which they belong. For each communication group, an objective and actions (improvement actions) for that objective are defined. Participants in the communication group post about improvement actions in the chat. Each employee belongs to one or more communication groups.

[0059] For example, the improvement action and objective of one social group is "stress reduction through the habit of morning walks." The improvement action and objective of another social group is "improvement of cognitive function and productivity through fast walking." The improvement action and objective of yet another social group is "prevention of mental illness through quality sleep."

[0060] The interaction data collection unit 19 periodically collects interaction data between employees in the interaction groups in which each employee participates from multiple employee terminals 3, analyzes the collected interaction data, and updates the interaction group DB 23 based on the analysis results. Chat data may be collected from a system that provides chat services. The interaction data to be collected includes information about each interaction group itself and chat data from each interaction group.

[0061] Each record in the interaction group DB23 represents information about a single interaction group. In the example configuration shown in Figure 5, the interaction group DB23 includes the improvement action column 231, the objective column 232, the number of group members column 233, the percentage of chat posters column 234, and the average number of chats per day column 235.

[0062] The interaction group DB23 further includes columns showing the number of employees in each employee classification within the interaction group. In Figure 5, columns 236 showing the number of employees in classification (a) and 237 showing the number of employees in classification (b) are shown as examples. The interaction data collection unit 19 can determine the number of employees in each classification from the estimation results of the employee classification estimation model 16 based on past interaction data.

[0063] The "Improvement Action" column 231 shows the actions that participants in the exchange group talk about in the chat. Each exchange group has a predefined improvement action and objective. For example, if the improvement action is "walking," participants post messages in the chat space indicating their walking results, helpful information related to walking, or responses to other people's posts. The "Objective" column 232 shows the objective of the exchange group. The "Number of Group Members" column 233 shows the number of people belonging to the group.

[0064] The "Chat Posting Percentage" column (357) shows the percentage of participants who posted a message during the specified period. Here, any participant who made at least one post is counted as a poster. Alternatively, only participants who posted more than the specified number of posts may be counted. The "Average Chat Frequency / Day" column (358) shows the average number of posts per day in the group during the specified period. The average unit of time may differ from a day.

[0065] The columns 233 (group size), 234 (percentage of chat posters), and 235 (average number of chats per day) each store the same type of data as the columns 356 (group size), 357 (percentage of chat posters), and 358 (average number of chats per day) in the training data 350 of the score improvement behavior estimation model 18.

[0066] The data in the Improvement Actions column 231 and the Objectives column 232 are fixed for each exchange group. The data in the other columns 233 to 237 are updated as needed according to the analysis results of the exchange data collection unit 19.

[0067] Figure 6 shows an example of the configuration of the interaction data DB 27. The interaction data DB 27 stores the features to be input to the employee classification estimation model 16 and the score estimation model 17. The interaction data collection unit 19 periodically collects interaction data between employees in the interaction groups in which each employee participates from multiple employee terminals 3, analyzes the collected interaction data to extract predetermined features, and stores the analysis results in the interaction data DB 27. The interaction data collection unit 19 also stores data indicating the content of chats in the interaction data DB 27 or in another database (not shown).

[0068] The interaction data DB27 contains items similar to the explanatory variables in the training data 310 for the employee classification estimation model and score estimation model shown in Figure 3. Each record shows information about a single employee's posts within a specific interaction group over a specific period.

[0069] Specifically, the interaction data DB27 includes the employee ID column 271 and the period column 272. Furthermore, it includes the number of posts column 273, the post sentiment analysis column 274, the number of predictive words column 275, the average post time column 276, the number of post responses column 277, and the interaction group column 278. These columns store the same type of data as the columns with the same names in the training data 310 for the employee classification estimation model and the score estimation model. The interaction group column 278 indicates the chat interaction group.

[0070] In the example in Figure 6, employee ID1 belongs to interaction group A and interaction group B, and information about the employee's posts in each interaction group is collected and stored periodically. Employee ID2 belongs to interaction group A, and information about the employee's posts in that group is collected and stored periodically. Old records may be deleted in a timely manner.

[0071] Figure 7 shows a flowchart of an example of the processing of the chat analysis device 1. In step S11, the communication data collection unit 19 collects employee communication data from each employee terminal 3. As explained with reference to Figures 5 and 6, the communication data collection unit 19 updates the communication group DB 23 and the communication data DB 27 based on the collected communication data. The update frequencies of the communication group DB 23 and the communication data DB 27 may differ.

[0072] The following describes the process for one employee within a single interaction group. The process described below may be performed for each employee within each interaction group.

[0073] In step S12, the employee classification estimation unit 10 uses the employee classification estimation model 16 to estimate the type of target employee in the target interaction group based on the data stored in the interaction data DB 27. The employee classification estimation unit 10 inputs the latest data of the target employee in the target interaction group into the employee classification estimation model 16 to determine the type of target employee.

[0074] Furthermore, in step S13, the stress and productivity score estimation unit 11 uses the score estimation model 17 to estimate the type of target employee in the target interaction group based on the data stored in the interaction data DB 27. The stress and productivity score estimation unit 11 inputs the data of the target employees in the target interaction group into the score estimation model 17 to calculate the stress score and productivity score of the target employees. For example, the productivity score and stress score are calculated based on the interaction data for a predetermined period immediately preceding the estimation of improvement behaviors.

[0075] Next, in step S14, the chat analysis device 1 calculates estimated fluctuations in the stress score and productivity score for each improvement action, based on the employee type, stress score, and productivity score.

[0076] First, the estimation result integration processing unit 12 generates data to be input to the score improvement behavior estimation model 18. Figure 8 shows an example of input data 400 to the score improvement behavior estimation model 18. The items of the input dataset shown in Figure 8 are the same as the items of the explanatory variables of the training data 350 of the score improvement behavior estimation model shown in Figure 4.

[0077] Specifically, the input data 400 consists of a classification column 402, a productivity score column 403 for action proposals, a stress score column 404 for action proposals, a proposed action column 405, a group size column 406, a chat post percentage column 407, and an average number of chats per day column 408. The data periods used to calculate the values ​​in different columns may be the same or different.

[0078] Each record represents a single input dataset. Multiple input datasets are sequentially input into the score improvement behavior estimation model 18, and stress score variability and productivity score variability are output. All records of the input data shown in Figure 8 represent combinations of information on the target employee in the target interaction group and different suggested behaviors.

[0079] The estimation result integration processing unit 12 stores the type of target employee estimated in step S12 in all cells of the classification column 402. The estimation result integration processing unit 12 obtains the productivity score and stress score estimated in step S13 and stores them in all cells of the action suggestion productivity score column 403 and the action suggestion stress score column 404. The estimation result integration processing unit 12 stores the pre-set different actions in the cells of the suggested action column 405.

[0080] Furthermore, the estimation result integration processing unit 12 retrieves the group size, chat posting percentage, and average number of chats per day for the target interaction group from the interaction group DB 23, using the group size column 233, the chat posting percentage column 234, and the average number of chats per day column 235. The estimation result integration processing unit 12 stores the retrieved values ​​in all cells of the group size column 406, the chat posting percentage column 407, and the average number of chats per day column 408, respectively.

[0081] Furthermore, the proposed action creation unit 13 sequentially inputs the input dataset (records) of the input data 401 into the score improvement action estimation model 18, and obtains estimated values ​​for the change in stress score and productivity score for each input dataset.

[0082] In one embodiment of this specification, the suggested action generation unit 13 may adjust the stress score and productivity score based on the employee's response to a predetermined number of suggested actions in the past (e.g., the previous time). This makes it possible to propose more appropriate action candidates. Note that the employee's response to the previous suggested action does not necessarily have to be referenced.

[0083] The Proposal Behavior Interest and Effectiveness Evaluation Unit 14 searches the chat data collected by the Interaction Data Collection Unit 19 for employee responses to the previous proposal. Responses can be obtained from "likes" posted to the proposal message and from messages posted after the proposal. The text analysis method is widely known, so details are omitted.

[0084] For example, if a "like" is posted or a related message is posted, the proposed behavior interest / effectiveness evaluation unit 14 may determine that there was interest or effectiveness, and may improve the stress score and productivity score of the same type of behavior as before by a predetermined value. If there is no response, or if a message rejecting the proposed behavior is posted, the proposed behavior interest / effectiveness evaluation unit 14 may determine that there was no interest or effectiveness, and may worsen the stress score and productivity score of the same type of behavior as before by a predetermined value. The correction values ​​may be changed according to the number of responses or the number of related messages posted. If information on the actual behavior of employees is available, the presence or absence of behavior in response to the suggestion may be referred to.

[0085] Next, in step S15, the chat analysis device 1 performs the processing from steps S12 to S14 for each of the other participants in the target interaction group. This provides an estimated change in stress score and an estimated change in productivity score for each of the other participants in the target interaction group.

[0086] The proposed action creation unit 13 stores the estimated changes in stress scores and productivity scores for all participants in the target interaction group into the improvement score table. Figure 9 shows an example of the structure of the improvement score table 450. In the example in Figure 9, employee ID 1 is the target employee, and employees ID 2 and ID 3 are the other participants in the target interaction group. Candidate improvement actions to be proposed are walking, sleeping, eating, participation in interaction group A, and participation in interaction group B.

[0087] In the next step, S16 and beyond, the proposed action creation unit 13 determines the type of improvement action to propose and the range of employees to whom it will be proposed. In step S16, the proposed action creation unit 13 determines the improvement action to propose to the target employee based on two score fluctuation values ​​of that employee. For example, if the two score fluctuation values ​​of the improvement action exceed their respective thresholds, it is determined that the improvement action should be presented to the target employee. Alternatively, the determination may be based on the sum of the two score fluctuation values ​​or a value entered into a predetermined function. There may be one or more improvement actions presented. For example, the improvement action estimated to be the best based on the two score fluctuation values ​​may be selected. Note that only one of the score fluctuation values ​​may be referenced.

[0088] Next, the proposed action creation unit 13 estimates the effective range of the improvement action selected for the target employee based on the statistical values ​​of the estimated score fluctuations of the improvement action. For example, the proposed action creation unit 13 calculates the mean and variance of the fluctuations in stress scores and productivity scores within the interaction group. The mean is an indicator of the improvement effect of the entire interaction group, and the variance is an indicator of the variability of the improvement effect within the interaction group.

[0089] The mean can be either a simple mean (with a common weight coefficient of 1) or a weighted mean, and the median may be used instead of the mean. The standard deviation or the maximum difference between the values ​​may be used instead of the variance.

[0090] The proposed action creation unit 13 compares the mean values ​​of the fluctuations in the stress score and productivity score with their respective thresholds, and further compares the variance values ​​of the fluctuations in the stress score and productivity score with their respective thresholds. If the mean values ​​of the fluctuations in the two scores exceed the threshold and their variance values ​​are less than the threshold (S16: YES), it is estimated that there is an effect on the entire interaction group. Otherwise (S16: NO), it is determined that there is little effect on the entire interaction group.

[0091] If it has an effect on the entire exchange group (S16:YES), in step S18, the proposed action creation unit 13 presents an improvement action in the group chat (S18). If it does not have an effect on the entire exchange group (S16:YES), in step S 17 In this step, the proposed action creation unit 13 presents improvement actions only to the target employee. Subsequently, the flow returns to step S13, where score re-estimation, score change value re-estimation, and revised proposal of improvement actions are performed.

[0092] In one embodiment of this specification, the proposed action creation unit 13 selects text to present to the employee from the improvement action DB 25. For example, the proposed action creation unit 13 selects text for the message from the improvement action DB 25 based on the type of improvement action decided to propose and the content of the message in the chat. The method for creating the message to be presented is known as chatbot technology, and details are omitted here.

[0093] The above example uses the score estimation model 17 to estimate the score for each type of improvement action. Other examples may input improvement actions that describe more specific content into the score estimation model 17. The input improvement actions may be templates that describe more specific content, or they may be messages that are actually presented to employees. If a template is input, a message to present may be selected from the improvement action DB 25 based on the template and the actual chat content. The proposed action creation unit 13 may also process messages stored in the improvement action DB 25 according to the chat content.

[0094] The following describes an example of a chat screen presented to an employee. Figures 10A to 10D show examples of screens displayed on all employee terminals 3 or on the employee terminal 3 of only one employee in a single communication group. The improvement action for the communication group in this example is "walk".

[0095] Figure 10A shows an example of messages in a chat room. Figure 10A shows messages from two people as an example. As described above, the chat analysis device 1 analyzes chat exchange data as illustrated in Figure 10A to estimate the effect of improvement behaviors on productivity and stress.

[0096] Figure 10B shows an example of displaying improvement actions proposed by the chat analysis device 1. In the example in Figure 10B, the chat analysis device 1 posts message 505 indicating an improvement action within the group chat. Improvement actions that are estimated to be effective for the entire exchange group can be shown to all participants in the exchange group.

[0097] Figure 10C shows another example of how improvement actions suggested by the chat analysis device 1 are displayed. In the example in Figure 10C, the chat analysis device 1 presents improvement actions on a suggestion screen dedicated to one participant. This allows improvement actions to be presented only to employees who need them. Participants can respond to the suggested action with a "like" 511 and input their actual step count 512.

[0098] Figure 10D shows an example of participant responses to improvement actions suggested in the chat room by the chat analysis device 1. In response to the improvement action suggestion message 505, a "Like" 521 and a related message 522 were returned. As described above, the chat analysis device 1 can correct the score fluctuation value based on the participant's responses 521 and 522 to the suggested improvement actions. Similarly, the chat analysis device 1 can correct the score fluctuation value based on the "Like" response 511 and the change in the actual number of steps 512 (change in actual behavior in response to the action suggestion), as shown in Figure 10C.

[0099] It should be noted that the present invention is not limited to the embodiments described above, and various modifications are included. For example, the embodiments described above are described in detail to make the present invention easier to understand, and are not necessarily limited to those having all the configurations described. Furthermore, it is possible to replace parts of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add configurations from other embodiments to the configuration of one embodiment. In addition, it is possible to add, delete, or replace parts of the configuration of each embodiment with other configurations. Furthermore, each of the above-mentioned configurations, functions, and processing units may be implemented in hardware, either partially or entirely, by designing them as integrated circuits, for example. Alternatively, each of the above-mentioned configurations and functions may be implemented in software by having the processor interpret and execute programs that implement each function. Information such as programs, tables, and files that implement each function can be stored in memory, a recording device such as a hard disk or SSD (Solid State Drive), or a recording medium such as an IC card or SD card. Furthermore, the control lines and information lines shown are those deemed necessary for explanatory purposes, and not all control lines and information lines are necessarily shown in the actual product. In practice, it is reasonable to assume that almost all components are interconnected. [Explanation of Symbols]

[0100] 1. Chat Analysis Device 3. Employee terminals 5 Network 10. Employee Classification Estimation Department 11. Stress and Productivity Score Estimation Unit 12. Estimation Result Integration Processing Unit 13. Proposal and Action Planning Department 14. Proposal Action Interest and Effectiveness Evaluation Department 15 Model Training Department 16. Employee Classification Estimation Model 17. Score Estimation Model 18. Estimation Model for Score Improvement Behavior 19. Exchange Data Collection Department 21 Training Data Database 23 Exchange Group Database 25 Improvement action DB 27 Exchange Data Database 201 CPU 202 memory 203 Auxiliary storage device 206 Communication devices

Claims

1. A device that proposes actions to participants in an exchange group, Processor and A storage device that stores the program to be executed by the aforementioned processor, Includes, The aforementioned processor, From the exchange data, which includes text messages posted by multiple participants in the first exchange group, the number of posts by the first participant and the number of words identical or similar to pre-set words are extracted for the first participant. At least a portion of the features of the first participant are input into a pre-configured score estimation model to calculate a predetermined score for the first participant. Based on at least a portion of the characteristics of the first participant, the personality type of the first participant is estimated. Based on the predetermined score and the type, estimate the change in the predetermined score for each candidate improvement action that improves the predetermined score. The candidate improvement actions include participation in an exchange group different from the first exchange group, A device that determines a proposed improvement action for the first participant based on the fluctuation value of the predetermined score.

2. The apparatus according to claim 1, The aforementioned processor, The information regarding the proposed improvement action is transmitted to the first participant's terminal, thereby proposing it to the first participant. After obtaining information on the first participant's response to the proposed improvement action, the predetermined score of the first participant is recalculated. Based on the recalculated predetermined score, the personality type of the first participant, and the response information, the change in the predetermined score is reestimated for each candidate improvement action that improves the predetermined score. A device that determines a revised improvement action to the first participant based on the re-estimated change in the predetermined score.

3. A device for suggesting actions to participants in an exchange group, Processor and A storage device that stores the program to be executed by the aforementioned processor, Includes, The aforementioned processor, From the exchange data, which includes text messages posted by multiple participants in the first exchange group, the characteristic quantities of each participant are extracted, including the number of posts by each participant and the number of words that are the same as or similar to pre-set words. At least a portion of the features of each of the multiple participants is input into a pre-configured score estimation model to calculate a predetermined score for each of the multiple participants. Based on the predetermined score of each of the multiple participants, the change in the predetermined score of each of the multiple participants is estimated for each candidate improvement action that improves the predetermined score. The candidate improvement actions include participation in an exchange group different from the first exchange group, A device that determines whether to present proposed improvement actions to the first exchange group or to individual participants based on the change in the predetermined score of each of the multiple participants.

4. The apparatus according to claim 1 or 3, The aforementioned processor, A device that estimates the change in the predetermined score for each candidate improvement action that improves the predetermined score, based on information from the first exchange group.

5. The apparatus according to claim 1 or 3, The device wherein the predetermined score is either a stress score or a productivity score.

6. The apparatus according to claim 5, The processor calculates a plurality of scores, including the predetermined score, The plurality of scores include the stress score and the productivity score, The processor is a device that determines a proposed improvement action based on the variation values ​​of the plurality of scores.

7. A method for proposing an action to a participant in an exchange group, The device extracts characteristic quantities of the first participant from the exchange data, which includes text messages posted by multiple participants of the first exchange group, including the number of posts by the first participant and the number of words that are identical or similar to pre-set words. The device inputs at least a portion of the features of the first participant into a pre-configured score estimation model to calculate a predetermined score for the first participant. Based on at least a portion of the characteristics of the first participant, the personality type of the first participant is estimated. The device estimates the change in the predetermined score based on the predetermined score and the type, for each candidate improvement action that improves the predetermined score. The candidate improvement actions include participation in an exchange group different from the first exchange group, A method wherein the apparatus determines a proposed improvement action for the first participant based on the fluctuation value of the predetermined score.

8. A method for proposing an action to a participant in an exchange group, The device extracts the characteristic quantities for each participant from the exchange data, which includes text messages posted by multiple participants in the first exchange group, including the number of posts by each participant and the number of words that are the same as or similar to pre-set words. The device inputs at least a portion of the feature quantities of each of the multiple participants into a pre-configured score estimation model to calculate a predetermined score for each of the multiple participants. The device estimates the change in the predetermined score of each of the multiple participants based on the predetermined score of each of the multiple participants, for each candidate improvement action that improves the predetermined score. The candidate improvement actions include participation in an exchange group different from the first exchange group, A method wherein the device determines whether to present a proposed improvement action to the first exchange group or to the individual participant, based on the change in the predetermined score of each of the multiple participants.

Images