Information processing device, method, and program

Abstract

The system determines the user's well-being status through scoring of the user's conversation content and supplementary processing based on data analysis, and provides a visual representation of the user's status. [Solution] The information processing device identifies question IDs and answer candidate scores from a set of conversation information read from the conversation log DB, merges the identified sets of question IDs and answer candidate scores, and stores the merged sets of question IDs and answer candidate scores in the well-being information DB. The information processing device determines which question IDs read from the question item DB are not stored in the well-being information DB, uses a plurality of answer groups stored in the memory unit to determine the score of the answer candidate associated with the question ID determined not to be stored in the well-being information DB, and provides the administrator terminal with visual information generated based on the question ID determined not to be stored and the score of the determined answer candidate.

Description

Technical Field

[0001] The present invention relates to an information processing apparatus, method, and program that determine the well-being state of a user through scoring of conversation content made by the user and complementary processing by data analysis, and visually provide the state of the user.

Background Art

[0002] In Japan, the decline in the working-age population due to the low birthrate and aging population and health problems is progressing. Along with this, problems such as intensified employment competition due to a shortage of human resources, the chronicity of long working hours, and an increase in the corporate burden of health insurance premiums have occurred, and companies are focusing on "health management (well-being)", which considers the health management of employees from a business perspective and implements it strategically. In health management, it is expected that making a health investment in employees based on the corporate philosophy will lead to the activation of the organization, such as an improvement in the vitality and productivity of employees, and ultimately to an improvement in performance and stock price. It is considered that improving the health status and psychological well-being of employees, interpersonal relationships within the organization, and career satisfaction will lead to an improvement in the overall quality of life of employees.

[0003] Citation Document 1 introduces a system for promoting the mental health care of each person throughout the workplace and constructing an environment in which employees can work with peace of mind in such a social situation.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] In recent years, companies have been promoting health management (well-being) to address the shortage of the workforce. Achieving health management requires understanding the employee situation, the environment of the workplace and care facilities, analyzing the results, and implementing countermeasures. Traditionally, companies considered employee health through regular health checkups, questionnaires, and interviews with supervisors. However, these measures made it difficult to grasp the employee's situation in a timely manner, and their implementation placed a significant burden on both supervisors and employees. Furthermore, interviews and questionnaires with supervisors failed to provide accurate information and insights into individual circumstances, such as health, dissatisfaction, and the causes of stress.

[0006] Furthermore, even when using technologies such as those disclosed in Reference 1, the amount of conversation between employees and robots varied from person to person, and not enough information was accumulated to accurately understand the employees' information and circumstances. For this reason, there was a growing need for a system to estimate the stress levels of employees and visualize their condition.

[0007] The present invention was made to solve these problems and aims to provide an information processing device, method, and program that determine the user's well-being state through scoring the content of conversations conducted by the user and supplementary processing through data analysis, and that visually provides the user's state. [Means for solving the problem]

[0008] An information processing apparatus according to one aspect of the present invention is an information processing apparatus comprising a communication unit, a storage unit, and a control unit for communicating with an external network, The aforementioned storage unit is A question item database that stores information on questions and possible answers for inquiring about the user's well-being status, A conversation log DB that stores a set of conversation information associated with the aforementioned user, A well-being information database that stores information on the user's well-being, Equipped with, The control unit, The system identifies the question ID and the score of the answer candidate from the set of conversation information read from the conversation log DB, and stores the identified question ID and the score of the answer candidate in the conversation log DB. The set of question IDs and answer candidate scores stored in the conversation log DB is merged, and the merged set of question IDs and answer candidate scores is stored in the well-being information DB. To determine which of the question IDs read from the aforementioned question item DB are not stored in the aforementioned well-being information DB, Using the multiple sets of answers stored in the memory unit, the score of the candidate answer associated with the question ID that is determined not to be stored in the well-being information DB is determined, and the question ID that is determined not to be stored in the well-being information DB and the score of the determined candidate answer are stored in the well-being information DB. The visual information generated based on the question ID and the score of the answer candidate stored in the well-being information DB is provided to the administrator terminal via the communication unit. It is configured to execute. [Effects of the Invention]

[0009] According to the present invention, it is possible to acquire the content of conversations between a user and a conversational device such as an autonomous robot, personal computer (PC), or smartphone. Therefore, it becomes possible to understand the user's well-being status from the content of the conversations the user has had, without having to conduct surveys or interviews with the user. [Brief explanation of the drawing]

[0010] A detailed understanding of the embodiments disclosed herein can be obtained from the following description illustrated in relation to the accompanying drawings. [Figure 1] This is a diagram showing the overall system configuration including the information processing device 10 according to the present invention. [Figure 2]It is a system configuration diagram of the information processing apparatus 10 according to the present invention. [Figure 3] It is a diagram for explaining an example of the data structure of the user data DB 106. [Figure 4] It is a diagram for explaining an example of the data structure of the question item DB 107. [Figure 5] It is a diagram for explaining an example of the data structure of the conversation log DB 108. [Figure 6] It is a diagram for explaining an example of the data structure of the well-being information DB 109. [Figure 7] It is a flowchart for explaining the conversation process with the user executed by the information processing apparatus 10 and the conversation apparatus 12. [Figure 8] It is a flowchart for explaining the process of estimating the score for each question item related to a predetermined well-being based on the data of the conversation log information 504, and the process of calculating by the average value of the scores of the answer candidates of some or all users. [Figure 9] It is a flowchart for explaining the process of generating the well-being score for each user and the process of generating the well-being level. [Figure 10] It is a diagram showing an example of the analysis result screen 1000 displayed on the administrator terminal 11.

Embodiment for Carrying Out the Invention

[0011] (Overall Configuration and Functions of Each Component) FIG. 1 is a configuration diagram of the entire system including the information processing apparatus 10 according to the present invention. The information processing apparatus 10 is connected to the administrator terminal 11 and the conversation apparatus 12 so as to be communicable with each other via wire or wireless. In FIG. 1, only one administrator terminal 11 and one conversation apparatus 12 are shown, but a plurality of these may exist.

[0012] In this specification, "user" means a user who receives the mental health support service described in this specification, for example, an employee belonging to a predetermined community such as a company, an individual in a nursing facility, a single person, and a delivery operator. The mental health support service includes providing a score associated with the well-being of each user obtained through a service for grasping the health status of the user and a service for watching over single persons. The well-being score is a numerical value indicating the physical, mental, and social states of the user, and in the examples described in this specification, it is assumed that the higher the score, the better the user's physical, mental, and social states. "Company" means not only a profit-making company but also any entity such as a local government or a non-profit organization. "Administrator" means a person who supports a user who receives the mental health support service, for example, a management position in a company or the like, an industrial physician, a counselor, a nursing facility staff member, a welfare provider, an operation manager, and the like.

[0013] The information processing device 10 can acquire environmental information of a specific area where the conversation device 12 circulates or is placed, for example, information such as temperature, humidity, illuminance, vibration, and wind acquired by various sensors, from the conversation device 12. The information processing device 10 can calculate the comfort level of the area environment such as a workplace based on the information acquired by various sensors.

[0014] The information processing device 10 can acquire the user ID or face image of the user with whom the conversation device 12 is attempting to have a conversation, perform an authentication process, and determine a conversation scenario based on at least one of the personal information of the authenticated user, the conversation log information of the user, the question items regarding well-being, and the environmental information of the specific area, and transmit a conversation sentence based on the conversation scenario, that is, voice data, to the conversation device 12.

[0015] For example, if a conversation scenario is based on the user's personal information, it may be a conversation scenario that asks about the user's hobbies and preferences; if it is based on the user's conversation log, it may be a conversation scenario related to what was said in past conversations; if it is based on well-being questions, it may be a scenario related to predetermined questions or questions that the user wants to ask the user in relation to those questions; and if it is based on environmental information of a specific area, it may be a conversation scenario related to general matters such as weather and temperature. A conversation scenario is information to confirm the user's situation, such as the user's mental health, and includes information that serves as a starting point (hint, beginning) for the conversation for the user. If information to confirm the user's situation can be obtained, the conversation may deviate from the conversation scenario. For example, if the conversation starts with talking about the weather and the user then talks about the difficulties of their job, the conversation may proceed to confirm the details of the difficulties of their job, without being limited to environmental topics such as the weather.

[0016] When the information processing device 10 determines a conversation scenario based on well-being-related questions, it can utilize information stored in the question item DB 107 for inquiring about the user's well-being status. This information may include, for example, questions from a 57-item questionnaire-style stress check test provided by the Ministry of Health, Labour and Welfare, or questions from employee engagement surveys conducted by various private companies, and is not particularly limited.

[0017] The information processing device 10 receives voice data (and optionally, a facial image) from the conversation device 12 and transmits voice data to the conversation device 12, thereby facilitating conversation between the conversation device 12 and the user. The information processing device 10 can analyze the voice data (and optionally, a facial image) received from the conversation device 12 to determine whether the user's conversation content is related to the conversation scenario and whether the user is interested in the conversation content. Based on these determinations, the information processing device 10 can control the subsequent conversation content.

[0018] The information processing device 10 can identify changes in the user over time based on the voice data and / or facial image received from the conversation device 12. For example, the information processing device 10 can read the user's conversation log information from the conversation log DB 108 and identify changes in the user's situation by identifying whether the user's statements are different from before, or whether the tone of their voice is higher or lower than a predetermined threshold, thereby controlling the content of subsequent conversations.

[0019] The information processing device 10 can generate candidate audio data for the next conversation to be conveyed to the user by transmitting a set of audio data from the conversation device 12 and the user's conversation, along with information on whether the user is interested in the conversation content, to the automatic conversation content generation unit. The automatic conversation content generation unit has a large-scale language model generated by machine learning and has the function of generating a response to the content of a question (prompt) in response to the receipt of the question. The question can be generated based on a set of audio data from the conversation device 12 and the user's conversation, along with a conversation scenario.

[0020] Machine learning may be performed by learning the results of a stress check test taken by a user in accordance with a 57-item questionnaire provided by the Ministry of Health, Labour and Welfare, or by learning the content of conversations between the user and the conversation device 12 as described herein, or by using other learning data (for example, survey data from employee engagement surveys conducted by various private companies).

[0021] The information processing device 10 can acquire a user's conversation log based on the user ID, and by sending the conversation log associated with the user and the answer to the question to the automatic conversation content generation unit, it can verify the content of the next conversation to be conveyed to the user and generate a revised candidate. The revised candidate becomes the audio data of the next conversation to be conveyed to the user. The information processing device 10 can generate a prompt based on the set of audio data of the conversation device 12 and the voice data of the voice spoken by the user, the candidate audio data of the conversation to be conveyed to the user following that set, the conversation scenario, and the user's conversation log.

[0022] The information processing device 10 can transmit the verified audio data of the next conversation to be conveyed to the user to the conversation device 12. By repeating the above process, the information processing device 10 can facilitate the conversation between the user and the conversation device 12.

[0023] The information processing device 10 can estimate the score of responses to well-being-related questions based on conversation log information, and can calculate the score of responses that cannot be estimated from the conversation log information using a predetermined method. The information processing device 10 can estimate the score of responses based on a natural language processing method. The information processing device 10 can calculate the correlation value between question items in each user's past response data, and can supplement the score of user responses that cannot be estimated from the conversation log information based on the calculated correlation value.

[0024] Furthermore, the information processing device 10 can calculate the average value of the question items in each user's past response data and, based on the calculated average value, can supplement the scores of users' responses that cannot be estimated from the conversation log information. If the supplemented score is not an integer, the information processing device 10 can round up or round down the score depending on the content of the well-being-related question items. Whether to use rounding up or rounding down is determined in a way that results in a relatively low-risk outcome, for example, to avoid missing highly stressed individuals. For example, when determining a user's work-related stress, if a higher response score indicates a higher stress level, rounding up will be performed.

[0025] The information processing device 10 can generate display data based on each user's data in the well-being information DB 109 and provide it to the administrator terminal 11. The display data may be, for example, a graph showing the time-series changes in each user's well-being score within the workplace, or a graph showing the scores for each of the indicators that make up the well-being score (e.g., job satisfaction, interpersonal stress, workload, etc.). The information processing device 10 can also compare the user's time-series data with that of people with similar past time-series data and display a warning (for example, if the current situation continues, the likelihood of leaving the company will increase) if a predetermined criterion is met.

[0026] The administrator terminal 11 may be any type of device capable of operating in a wired or wireless environment (e.g., a smartphone, PC, tablet, etc.) and is not limited to any specific device. The administrator terminal 11 communicates with the information processing device 10 and sends and receives various information. The administrator terminal 11 can send user data to be entered as master data, along with questions and answer candidates for inquiring about the user's well-being status, to the information processing device 10, and receive the analysis results from the information processing device 10. Therefore, the analysis results will be provided to those who have a role in supporting the user's mental health.

[0027] The conversation device 12 may be an autonomous robot, a PC, or a smartphone. The conversation device 12 may be an autonomous robot equipped with various sensors and based on SLAM (Simultaneous Localization and Mapping) technology, capable of patrolling a specific area such as a company workplace. The conversation device 12 may be a PC or smartphone equipped with various sensors, capable of being located near the user. The conversation device 12 can transmit sensor data about the environment of the user's area, such as temperature, humidity, illuminance, vibration, and wind, acquired by the various sensors, to the information processing device 10. The information processing device 10 stores the received sensor data as environmental information and can use this environmental information for conversations with the user.

[0028] The conversation device 12 may include short-range wireless communication means, communication means, a speaker, a microphone, a camera, and a display. The conversation device 12 can output voice based on voice data received from the information processing device 10 through its speaker, or display text messages based on voice data on its display. The conversation device 12 can convert the user's voice acquired by the microphone into voice data and transmit it to the information processing device 10, and can also transmit the user's image data acquired by the camera to the information processing device 10.

[0029] The conversation device 12 can transmit a user ID provided by the user to the information processing device 10 for user authentication processing, and / or can acquire a facial image of the user with the camera and transmit the acquired facial image to the information processing device 10 for user authentication processing.

[0030] The conversation device 12 can communicate with the information processing device 10 via wired or wireless communication means and send and receive various types of information. The conversation device 12 can acquire information such as the name, personal information, and voice data based on a conversation scenario of a user authenticated by the information processing device 10, and can start a conversation with that user. The conversation device 12 recognizes the user's voice received via the microphone, generates voice data based on the recognized voice and transmits it to the information processing device 10, and then emits synthesized voice from the speaker based on the voice data received from the information processing device 10, thereby engaging in conversation with the user.

[0031] The conversation device 12 transmits conversation information input by the user to the information processing device 10, and can display the audio data received from the information processing device 10 as text on the display. This allows the user to know the content of their own conversation and the content of the audio emitted by the conversation device 12 as text information. The conversation device 12 can also transmit image information of the user acquired by the camera to the information processing device 10 for the purpose of determining the user's feelings, which is performed by the information processing device 10. The image information may include not only facial images, but also images of other parts of the body, or even the entire body.

[0032] (System Configuration) Figure 2 is a system configuration diagram of the information processing device 10 according to the present invention. The information processing device 10 comprises a control unit 101, a main memory unit 102, an auxiliary memory unit 103, an IF unit 104, and an output unit 105, which are interconnected by a bus 120 or the like, similar to a general computer. The information processing device 10 includes a user data DB 106, a question item DB 107, a conversation log DB 108, and a well-being information DB 109 in the form of storage means such as files / databases within the auxiliary memory unit 103.

[0033] The control unit 101, also known as the central processing unit (CPU), controls each component of the information processing device 10 and performs data calculations. It also reads various programs stored in the auxiliary storage unit 103 into the main memory unit 102 and executes them. The main memory unit 102, also known as main memory, can store various received data, computer-executable instructions, and data after calculations performed by those instructions. The auxiliary storage unit 103 is a storage device such as a hard disk drive (HDD) or solid-state drive (SSD), and is used for long-term storage of data and programs.

[0034] The embodiment shown in Figure 2 describes an embodiment in which the control unit 101, main memory unit 102, and auxiliary storage unit 103 are located within the same computer. However, in other embodiments, the information processing device 10 can be configured to achieve parallel distributed processing by multiple computers by using multiple control units 101, main memory unit 102, and auxiliary storage unit 103. Furthermore, in other embodiments, it is also possible to set up multiple servers for the information processing device 10, and have multiple servers share a single auxiliary storage unit 103.

[0035] The IF unit 104 acts as an interface (IF) for sending and receiving data with other systems and devices, and also provides an interface for receiving various commands and input data (various masters, tables, etc.) from the system operator. The output unit 105 provides a display screen for displaying the processed data and printing means for printing the data.

[0036] Similar functional components to the control unit 101, main memory unit 102, auxiliary memory unit 103, IF unit 104, and output unit 105 also exist in the administrator terminal 11, but their description is omitted in this specification.

[0037] The user data DB 106 stores information about users receiving the mental health support services described herein. Figure 3 illustrates an example of the data structure of the user data DB 106. The user data DB 106 may include, but is not limited to, user ID 301, user information 302, and user image 303, and may also include other data items. For example, it may be configured to include an identifier that identifies a community to distinguish employees of different companies (e.g., a company ID). The user data DB 106 may also be managed by separating storage means such as files / databases for each community.

[0038] User ID 301 is an identifier that identifies the user. User information 302 shows information about the user, such as the user's name, affiliated organization, position, occupation, gender, date of birth, and hobbies. User image 303 includes at least a facial image of the user and may include a full-body image.

[0039] Returning to Figure 2 and continuing the explanation, the Question Item DB107 stores information on well-being-related questions, specifically, the content of questions and candidate answers for inquiring about the user's well-being status. The candidate answer information includes a score (points) as part of it. Figure 4 illustrates an example of the data structure of the Question Item DB107. The Question Item DB107 can include Question ID 401, Question Content 402, and Candidate Answer 403, but is not limited to these data items and can include other data items as well.

[0040] Question ID 401 indicates information that identifies the question item, and may be information such as "Q1" or "Q2". Question content 402 is a question to ask about the user's well-being state, and is not limited to that, but may be a question item like the 57-item questionnaire-style stress check test provided by the Ministry of Health, Labour and Welfare, or a question item like those used in employee engagement surveys conducted by various private companies. Answer candidates 403 may be equivalent to answer candidates in a stress check test, and show multiple answer candidates for the question in Question Content 402 (for example, "1: Satisfied, 2: Somewhat satisfied, 3: Somewhat dissatisfied, 4: Dissatisfied", "1: Yes, 2: Somewhat yes, 3: Somewhat different, 4: Different", etc.). Answer candidates 403 may include a score as a value for each answer candidate.

[0041] Returning to Figure 2 and continuing the explanation, the conversation log DB 108 stores the content of the conversation between the user and the conversation device 12. Figure 5 is a diagram illustrating an example of the data structure of the conversation log DB 108. The conversation log DB 108 can include user ID 301, conversation ID 501, conversation start time 502, conversation end time 503, conversation log information 504, and well-being score information 505, but it is not limited to these data items and can include other data items as well.

[0042] User ID 301 is an identifier that identifies the user. Conversation ID 501 is an identifier that identifies a set of conversation information. Conversation start time 502 indicates the time when the user and the conversation device 12 started the conversation. Conversation end time 503 indicates the time when the user and the conversation device 12 ended the conversation. This time information includes the year, month, day, hour, minute, and second. Conversation log information 504 indicates one or more sets of conversation content between the user and the conversation device 12. Wellbeing score information 505 includes question IDs and answer candidate information associated with each of the conversation contents.

[0043] Returning to Figure 2 and continuing the explanation, the Wellbeing Information DB 109 stores information on each user's wellbeing level each month, as well as detailed information that constitutes the wellbeing level. Figure 6 is a diagram illustrating an example of the data structure of the Wellbeing Information DB 109. The Wellbeing Information DB 109 can include user ID 301, target year and month 601, basic data 602, and wellbeing level 603, but is not limited to these data items and can include other data items as well.

[0044] User ID 301 is an identifier that identifies the user. Target year and month 601 indicates the year and month information of the well-being level. Basic data 602 indicates the basic data for generating data that shows the well-being status associated with the user. Well-being level 603 indicates data that shows the well-being status generated based on basic data 602.

[0045] (Control flow of conversation processing by information processing device 10 and conversation device 12) Figure 7 is a flowchart illustrating the user conversation process performed by the information processing device 10 and the conversation device 12. Before starting the conversation, the information processing device 10 completes user authentication by querying the user data DB 106 based on the user ID and user facial image data received from the conversation device 12.

[0046] In S701, the information processing device 10 determines a conversation scenario based on at least one of the following: well-being-related questions, the authenticated user's personal information, the user's conversation log information, information obtained by analyzing the user's voice data and / or facial image received from the conversation device 12, and environmental information of the area where the user is located. Based on the determined conversation scenario, the device initiates a conversation with the user.

[0047] In S702, the information processing device 10 queries the question item DB 107 to retrieve information for an arbitrary question ID 401 and question content 402. The information processing device 10 can generate a conversation text based on the retrieved question content 402 information and at least one of the following: the user's personal information, the user's conversation log information, information obtained by analyzing the user's voice data and / or facial image received from the conversation device 12, and environmental information of the area where the user is located.

[0048] The generation of conversational text may be performed by the automatic conversational content generation unit of the information processing device 10. The automatic conversational content generation unit has a large-scale language model generated by machine learning and has the function of generating conversational text in response to the content of a question (prompt) received. The information processing device 10 can also specify the speaker's personality to the automatic conversational content generation unit, for example, it can be set to play the role of an experienced counselor. The automatic conversational content generation unit may be located in a device connected to an external network or in the information processing device itself.

[0049] The generated conversation text may be rephrased based on the user's personal information or conversation log information, for example, so that the content of question 402 can be asked in a way that suits the user's situation. The generated conversation text may also include the value of question ID 401 associated with the read question 402 (for example, "Q1"). The generated conversation text may be, for example, a phrase asking about the user's workload, such as "It seems like your overtime hours have increased recently, are you feeling alright?" or a phrase asking about how they spend time with their family, such as "You said you were going to visit XX last weekend, how did it go?" The information processing device 10 stores the generated conversation text in the conversation log information 504 of the conversation log DB 108.

[0050] When generating conversational text, the information processing device 10 can define relevant keywords from the first and second answer groups described later, referring to Figure 8, and generate conversational text for questions in a way that makes it easier for the user's answers to include the keywords. If the keywords are included in the user's answers, it becomes easier to estimate the user's answers to other question IDs, even if no answer candidates for other question IDs are included in the conversation log, resulting in a more accurate final estimation result.

[0051] Optionally, the information processing device 10 may have a function to verify the generated conversation text. The information processing device 10 can query the user data DB 106 and the conversation log DB 108 to have the conversation content automatic generation unit verify the validity of the generated conversation text, and can also regenerate the conversation text based on the verification results.

[0052] In S703, the information processing device 10 transmits the generated conversation text to the conversation device 12 as audio data for use by the conversation device 12. The conversation device 12 synthesizes speech based on the received audio data and emits the speech to the user through its speaker.

[0053] In S704, the conversation device 12 receives the audio of the user's conversation and transmits the generated audio data (and optionally the user's face image acquired by the camera) based on the received audio to the information processing device 10. The information processing device 10 queries the question item DB 107 based on the value of the question ID 401 associated with the conversation text generated in S702 and obtains information on the answer candidate 403. The information processing device 10 analyzes the user's audio data (and optionally the face image) received from the conversation device 12, determines which of the obtained answer candidate 403 information corresponds to the analysis result, associates the information of the determined answer candidate with the audio data of the user's conversation, and stores it in the conversation log information 504 of the conversation log DB 108 as information set with the conversation text generated in S702. Through this process, information on which question content each set of conversation between the user and the conversation device 12 corresponds to is stored in the conversation log DB 108. As mentioned above, since each answer option can include a score as a value, a set of conversations can include a question ID (e.g., Q1) and a score for each answer option (e.g., "4").

[0054] In S705, the information processing device 10 analyzes the user's voice data and / or facial image data received via the conversation device 12 to determine whether the user intends to continue the conversation. This determination may be based on whether the voice data contains positive or negative information, or it may be made by estimating the user's emotions from the user's facial image. If it is determined that the user intends to continue the conversation, the process returns to S702, and the information processing device 10 can generate the subsequent conversation sentence. On the other hand, if it is determined that the user does not intend to continue, the information processing device 10 terminates this processing flow.

[0055] By repeating the process described above, the conversation between the user and the conversation device 12 continues, and the conversation log information 504 in the conversation log DB 108 stores the set of conversations that took place between the user and the conversation device 12. Even if the conversation scenario is a set of conversations that is not based on well-being-related questions, the process in S801 described later will enable the set of conversations to include question IDs and scores for answer candidates. As a result, the information processing device 10 can provide mental health support services even for information on conversations between users that do not involve the conversation device 12.

[0056] (Pre-processing flow before providing mental health support services) Figure 8 is a flowchart illustrating the process of estimating a score for each of the predetermined well-being-related question items based on the data from the conversation log information 504, and the process of calculating the score using the average of the scores of some or all of the user's answer candidates.

[0057] In S801, the information processing device 10 accesses the conversation log DB 108 and reads the conversation log information 504 data for some or all users. When reading some of the data, one or more of the user's organization, occupation, age group, and gender may be used as search conditions. The information processing device 10 determines whether the set of conversations between the user and the conversation device 12 contained in each of the read conversation log information data contains the value of question ID 401 and the answer candidate information. If it is determined that it does, the process proceeds to S802.

[0058] On the other hand, if it is determined that the question is not included, the information processing device 10 queries the question item DB 107 to read the information of the question content 402, analyzes the conversation text in the set of conversations to determine which question content it corresponds to, and determines the value of the corresponding question ID based on the result of the determination. Furthermore, the information processing device 10 reads the information of the answer candidates associated with the determined question ID value, analyzes the conversation text in the set of conversations to determine which answer candidate it corresponds to, and determines the corresponding answer candidate based on the result of the determination. The information processing device 10 includes the determined question ID (e.g., Q1) and the score for the answer candidate (e.g., "4") in the set of conversations and stores the included data in the conversation log DB 108. As a result of this process, each set of conversations included in the data of the conversation log information 504 will include the question ID and the score for the answer candidate. Then, the process proceeds to S802.

[0059] In S802, the information processing device 10 reads conversation log data 504 for some or all users from the conversation log DB 108. The information processing device 10 extracts information on question IDs and scores for answer candidates from the set of conversation information between the user and the conversation device 12 contained in the read data. As described above, the set of conversation information after processing in S801 includes question IDs and scores for answer candidates. Based on the extracted information on question IDs and scores for answer candidates, the information processing device 10 calculates a first correlation coefficient between question IDs.

[0060] In S803, the information processing device 10 determines a first threshold from the first correlation coefficient calculated based on predetermined criteria. The first threshold may be, for example, a correlation coefficient of ±0.5 or greater, but is not particularly limited.

[0061] In S804, the information processing device 10 can determine a group of question IDs (first answer group) based on whether the calculated first correlation coefficient satisfies the first threshold. For example, if the question IDs are Q1 to Qn (where n is an integer representing the number of questions), they can be grouped as follows. For the sake of explanation, information on the score of the answer candidates is not shown within the group, but each question ID is associated with information on the score of the corresponding answer candidate. In this example, it is shown that because the content of questions Q1 and Q4 are similar, the answers to the questions tend to have similar tendencies. Group 1: Q1, Q4 Group 2: Q6, Q8, Q9, Q12, Q15 Group 3: Q16, Q17, Q18, Q21, Q25, Q26] ... Group m: Q47, Q52, Q54, Q57 In the example above, for "Group 1," it can be determined that Q1 and Q4 will follow the same response trend. Therefore, as will be explained later with reference to Figure 9, if the candidate answers for Q1 are known, even if the data for Q4 is not present in the user's conversation set, it becomes possible to estimate candidate answers for Q4 that follow the same response trend as Q1.

[0062] In S805, the information processing device 10 reads conversation log data 504 for some or all users from the conversation log DB 108. The information processing device 10 extracts information on question IDs and scores for answer candidates from the set of conversation information between the user and the conversation device 12 contained in the read data. As described above, the set of conversation information after processing in S801 includes question IDs and scores for answer candidates. The information processing device 10 calculates a second correlation coefficient between answer candidates for any n questions (where n is an integer). For example, if Q1 and Q2 are selected as n questions, the second correlation coefficient is calculated by determining which answer candidate is selected most often in Q2 when answer candidate "1" is selected in Q1, the percentage of times answer candidate "1" is selected in Q2, the percentage of times answer candidate "2" is selected in Q2, the percentage of times answer candidate "3" is selected in Q2, and the percentage of times answer candidate "4" is selected in Q2.

[0063] In S806, the information processing device 10 determines a second threshold from the second correlation coefficient calculated based on predetermined criteria. The second threshold may be, for example, a correlation coefficient of ±0.5 or greater, but is not particularly limited.

[0064] In S807, the information processing device 10 can determine a combination of question ID and candidate answer scores (second answer group) based on whether the calculated second correlation coefficient satisfies the second threshold. The second answer group may be, for example, as follows. Q1-“1”:Q4-“4”, Q7-“2”, Q8-“3” Q1-“2”:Q3-“1”, Q4-“3”, Q7-“1”, Q10-“1” ... Q50-“4”: Q55-“1”, Q56-“1”, Q57-“4”, Q58-“4” In the example above, it can be concluded that if a user selects "1" as their answer to Q1, they are more likely to answer "4" to Q4. Therefore, as will be explained later with reference to Figure 9, if we know that the answer candidate for Q1 is "1", we can estimate that the answer candidate for Q4 is "4" even if the data for Q4 is not present in the user's conversation set.

[0065] As described above, each of the answer candidates included in the first and second answer groups has a correlation coefficient. Therefore, when multiple answer candidates are available, the system is configured to select the answer candidate with the relatively highest correlation coefficient.

[0066] In S808, the information processing device 10 queries the question item DB 107 to read all combinations of question ID and answer candidate scores, and identifies combinations from the read combinations in which the first correlation coefficient does not meet the first threshold and the second correlation coefficient does not meet the second threshold. The information processing device 10 reads conversation log information 504 data for some or all users from the conversation log DB 108 that matches the identified question ID and answer candidate score combination. When reading some of the data, one or more of the user's organization, occupation, age group, and gender may be used as conditions.

[0067] The information processing device 10 extracts information on question IDs and answer candidate scores included in the set of conversational information read out, and calculates the average score of the answer candidate for each extracted question ID. For example, if the answer candidates corresponding to an arbitrary question ID are "1: Satisfied, 2: Somewhat satisfied, 3: Somewhat dissatisfied, 4: Dissatisfied", the number of "1: Satisfied" responses is calculated (10), the number of "2: Somewhat satisfied" responses is calculated (35), the number of "3: Somewhat dissatisfied" responses is calculated (18), and the number of "4: Dissatisfied" responses is calculated (7). Based on the calculated number of responses, the average value is calculated (10 × 1 + 35 × 2 + 18 × 3 + 7 × 4) / 70 = 2.314...).

[0068] The information processing device 10 can determine a third set of answers by rounding up or down decimal places based on predetermined criteria for each question ID. The choice between rounding up or rounding down is based on which option would increase the user's stress level. In the example above, for instance, if the question was "Are you satisfied with your relationships at work?", it can be determined that rounding up "2.314..." to "3: Somewhat dissatisfied" would result in a higher stress level, so rounding up is performed for such questions. Therefore, whether to round up or round down the calculated average can be determined on a question-by-question basis.

[0069] In S809, the information processing device 10 stores the first answer group, the second answer group, and the third answer group in the auxiliary storage unit 103. In the example described herein, an embodiment is described in which the first answer group, the second answer group, and the third answer group are generated each time the process shown in Figure 9, which is described below, is performed, but the embodiments of the present invention are not limited thereto. For example, the system may be configured to generate the first answer group, the second answer group, and the third answer group on a predetermined day each month.

[0070] (Wellbeing score generation process and wellbeing level generation process flow) Figure 9 is a flowchart illustrating the process of generating a well-being score and a well-being level for each user. In this embodiment, this process is performed once a month.

[0071] In S901, the information processing device 10 reads data for a given user from the conversation log DB 108 one record at a time for a specified year and month, based on the year, month, day, hour, minute, and second information of the conversation start time 502. The information processing device 10 identifies the question ID and the set of answer candidate scores from the set of conversation information information contained in the conversation log information 504 of the read data.

[0072] Table 1 shows examples of identified question IDs and answer candidate scores. For example, if the answer candidates are information such as "1: Satisfied, 2: Somewhat satisfied, 3: Somewhat dissatisfied, 4: Dissatisfied," then each candidate includes a numerical value, which is shown in Table 1. The information processing device 10 stores the set of identified question IDs and answer candidate scores in the well-being score information 505 of the read data.

[0073] [Table 1]

[0074] The information processing device 10 performs the above-described process for all data of the user for a predetermined year and month, and stores the identified question ID and set of answer candidate scores in the well-being score information 505 for each data. As a result of this process, the well-being score information 505 for all data of the user for a predetermined year and month will include all the information of the question ID and set of answer candidate scores identified from the conversation information set.

[0075] The information processing device 10 reads the well-being score information 505 for a specified month and year of the user from the conversation log DB 108, extracts a set of question IDs and answer candidate scores from the read well-being score information 505, and merges the extracted sets. If there are duplicate question IDs among the merged sets of question IDs and answer candidates, the information processing device 10 calculates the average value of the answer candidate values. If the calculated average value is not an integer, the information processing device 10 rounds up or down the decimal part based on a predetermined criterion for each question ID, in the same manner as described above.

[0076] The information processing device 10 adds the merged and (optionally) rounded-up / rounded-down sets of question IDs and answer candidate scores to the base data 602 of the well-being information DB 109. At this point, the well-being information DB 109 will have the base data for any given user's target year and month (i.e., the predetermined year and month that was processed), but the number of question ID and answer candidate score sets added to the base data 602 may be less than the number of question IDs and answer candidates stored in the question item DB 107. In such cases, the processing described below is executed to supplement the sets of question IDs and answer candidate scores that do not exist in the base data 602.

[0077] In S902, the information processing device 10 reads the first answer group, the second answer group, and the third answer group from the auxiliary storage unit 103.

[0078] In S903, the information processing device 10 accesses the well-being information DB 109 and reads the data for the user in question for the target year and month from the basic data 602. The information processing device 10 reads all question IDs (Q1, Q2, ... Qn) from the question item DB 107 and determines which of the read question IDs are not included in the data of the basic data 602 (for example, Q8 and Q9 are not included).

[0079] The information processing device 10 determines which group of the first answer group the determined question ID belongs to and identifies other question IDs (e.g., Q6) included in that group. The information processing device 10 identifies the score of the answer candidate associated with the identified question ID from the first answer group, and determines that for the answer candidate of the question ID that was determined not to be included in the data of the basic data 602 above, the answer candidate with the same tendency as the identified answer candidate has been selected. The information processing device 10 stores the determined set of question ID and answer candidate score in the basic data 602. In this specification, the processing of S903 using the first answer group will be referred to as the "first supplementary processing".

[0080] In S904, the information processing device 10 accesses the well-being information DB 109 and reads the data for the user in question for the target year and month from the basic data 602. The information processing device 10 reads all question IDs (Q1, Q2, ... Qn) from the question item DB 107 and determines which of the read question IDs are not included in the data of the basic data 602 (for example, Q50 is not included).

[0081] The information processing device 10 determines which group of the second set of answers the determined question ID belongs to, and identifies other question IDs (e.g., Q55, Q56, Q57, Q58) that are included in that group. The information processing device 10 determines whether the identified question ID is included in the data of the basic data 602, and extracts sets of question IDs and answer candidate scores that it determines are included (e.g., Q55-"1", Q56-"1", Q57-"4", Q58-"4"). The information processing device 10 determines whether the extracted sets of question IDs and answer candidates are included in the data of the basic data 602, and identifies sets of question IDs and answer candidates that it determines are included (e.g., Q57-"4"). Based on the identified sets of question IDs and answer candidates (e.g., Q57-"4"), the information processing device 10 estimates answer candidates for question IDs (e.g., Q50) that are not included in the data of the basic data 602, and determines that these are sets of question IDs and answer candidates that should be supplemented. The information processing device 10 stores the determined question ID and set of answer candidates in the basic data 602. In this specification, the processing in S904 using the second set of answers will be referred to as the "second supplementary processing".

[0082] In S905, the information processing device 10 accesses the well-being information DB 109 and reads the data for the user in question for the target year and month from the basic data 602. The information processing device 10 reads all question IDs (Q1, Q2, ... Qn) from the question item DB 107 and determines which of the read question IDs are not included in the data of the basic data 602 (for example, Q51 is not included).

[0083] The information processing device 10 searches for a third set of answers based on the determined question ID, identifies the score of the answer candidate associated with the question ID, and determines that this set of question ID and the identified answer candidate score is the set of question ID and answer candidate that should be supplemented. The information processing device 10 stores the determined set of question ID and answer candidate in the basic data 602. In this specification, the processing in S905 using the third set of answers will be referred to as the "third supplementation processing".

[0084] As a result of the above processing, the base data 602 will include information on all question IDs from the question item DB 107. Even if no actual conversation takes place between the user and the conversation device 12, the data can be estimated and supplemented by processing S903 to S905. The order in which processing S903 and S904 are performed does not matter and is not particularly limited.

[0085] In S906, the information processing device 10 calculates the sum of the score of the answer candidates based on the information of the set of question IDs and answer candidate scores contained in the basic data 602, and stores it in the well-being level 603.

[0086] In S907, the information processing device 10 queries the well-being information DB 109, generates data to be displayed on the administrator terminal 11 based on the basic data 602 and well-being level 603 information of the retrieved data, and can send the generated data to the administrator terminal 11.

[0087] Figure 10 shows an example of the analysis results screen 1000 displayed on the administrator terminal 11. The analysis results screen 1000 displays a graph showing the time-series change in the well-being level of users belonging to a certain workplace over a predetermined period, the results of the well-being level assessment for each employee in the latest month (for example, Alice's level has significantly declined), and a graph showing the details of the stress level of a specific user (Alice). In the example of the graph showing the details of the stress level on the analysis results screen 1000, it is shown that the user (Alice) is experiencing stress in terms of the quality of work and the work environment, mental and physical anxiety, and support from superiors and colleagues. In this way, the analysis results screen 1000 can show the well-being level and detailed information for each user, categorized by the organization to which the user belongs, occupation, age, and gender. The information processing device 10 can generate the analysis results screen 1000 based on the information stored in the well-being information DB 109 and provide it to the administrator terminal 11.

[0088] Furthermore, the information processing device 10 can compare the time-series changes in a user's well-being level over a predetermined period with the time-series changes of past users who have left the company or developed illnesses, and can display an alarm to the administrator terminal 11 if similar signs are observed.

[0089] Although the principles of the present invention have been described above with reference to exemplary embodiments, those skilled in the art will understand that various embodiments with modifications in configuration and details can be realized without departing from the spirit of the invention. That is, the present invention can take the form of, for example, a system, apparatus, method, program, or storage medium. [Explanation of Symbols]

[0090] 10 Information Processing Devices 11 Administrator terminal 12 Conversation device 101 Control Unit 102 Main memory 103 Auxiliary storage 104 IF section 105 Output section 106 User Data DB 107 Question item DB 108 Conversation Log Database 109 Wellbeing Information Database

Claims

1. An information processing device comprising a communication unit, a storage unit, and a control unit for communicating with an external network, The aforementioned storage unit is A question item database that stores information on questions and possible answers for inquiring about the user's well-being status, A conversation log DB that stores a set of conversation information associated with the aforementioned user, A well-being information database that stores information on the user's well-being, Equipped with, The control unit, The system identifies the question ID and the score of the answer candidate from the set of conversation information read from the conversation log DB, and stores the identified question ID and the score of the answer candidate in the conversation log DB. The set of question IDs and answer candidate scores stored in the conversation log DB is merged, and the merged set of question IDs and answer candidate scores is stored in the well-being information DB. To determine which of the question IDs read from the aforementioned question item DB are not stored in the aforementioned well-being information DB, Using the multiple sets of answers stored in the memory unit, the score of the candidate answer associated with the question ID that is determined not to be stored in the well-being information DB is determined, and the question ID that is determined not to be stored in the well-being information DB and the score of the determined candidate answer are stored in the well-being information DB. The visual information generated based on the score of the question ID and the candidate answer stored in the well-being information DB is provided to the administrator terminal via the communication unit. An information processing device configured to perform the following actions.

2. The aforementioned multiple response groups include a first response group and a second response group, The control unit, Extracting the question ID and answer candidate scores included in the set of conversation information for some or all users read from the conversation log DB, Based on the extracted question IDs and score information for the answer candidates, a first correlation coefficient between the question IDs is calculated, and a first set of answers is determined based on the first correlation coefficient that satisfies a first threshold. Based on the extracted question IDs and score information for the answer candidates, a second correlation coefficient is calculated between the answer candidates for n questions (where n is an integer), and a second set of answers is determined based on the second correlation coefficient that satisfies the second threshold. The information processing apparatus according to claim 1, configured to further perform the following:

3. The aforementioned multiple response groups include a third response group, The control unit, This involves reading all combinations of the question ID and the score of the answer candidate from the aforementioned question item database, Identifying combinations from the read-out combinations in which the first correlation coefficient does not satisfy the first threshold and the second correlation coefficient does not satisfy the second threshold, Retrieving from the conversation log DB a set of conversation information for some or all users that matches the identified combination of question ID and answer candidate score, The system extracts the question ID and answer candidate score information from the set of conversation information read out, and calculates the average score of the answer candidates for each extracted question ID. The third set of answers is determined by rounding up or rounding down the decimal part of the average value based on predetermined criteria for each question ID. The information processing apparatus according to claim 2, configured to further perform the following:

4. The information processing apparatus according to claim 3, wherein the selection of the rounding up process or the rounding down process is made based on which option would increase the user's stress level.

5. The control unit, Associating the question ID with the first conversation sentence of the first voice data to be transmitted via the communication unit to the conversation device that converses with the user, Associating the score of the answer candidate associated with the question ID with the second conversation sentence based on the audio data received from the conversation device, The first conversation sentence and the second conversation sentence are stored in the conversation log DB as a set of conversation information associated with the user, The information processing apparatus according to claim 1, configured to further perform the following:

6. The information processing apparatus according to claim 5, wherein the control unit is configured to generate the first conversation sentence based on keywords associated with at least a portion of the plurality of answer groups.

7. A method performed by an information processing device comprising a communication unit, a storage unit, and a control unit for communicating with an external network, The aforementioned storage unit is A question item database that stores information on questions and possible answers for inquiring about the user's well-being status, A conversation log DB that stores a set of conversation information associated with the aforementioned user, A well-being information database that stores information on the user's well-being, Equipped with, The control unit identifies the question ID and the score of the answer candidate from the set of conversation information read from the conversation log DB, and stores the identified question ID and the score of the answer candidate in the conversation log DB. The control unit merges the sets of question IDs and answer candidate scores stored in the conversation log DB and stores the merged sets of question IDs and answer candidate scores in the well-being information DB. The control unit determines which of the question IDs read from the question item DB are not stored in the well-being information DB. The control unit uses a plurality of answer groups stored in the storage unit to determine the score of the answer candidate associated with the question ID that has been determined not to be stored in the well-being information DB, and stores the question ID that has been determined not to be stored in the well-being information DB and the score of the determined answer candidate in the well-being information DB. The control unit provides the administrator terminal with visual information generated based on the score of the question ID and the answer candidate stored in the well-being information DB, via the communication unit. A method for providing this.

8. A program that causes a computer to perform the method described in claim 7.

Images