Information processing apparatus and method, and information processing system

Abstract

To provide an information processor, an information processing method and an information processing system for accurately estimating the state of the wellbeing of a user, and for presenting proper advice for improving the wellbeing to the user.SOLUTION: A method executed by a well-being monitor apparatus (information processing apparatus) that inputs time-series sensor measurement data output from an acceleration sensor of a wearable terminal worn by a user 6 to a predetermined model and estimates a transition of an emotional state constituting a well-being state, the method comprising: A model is periodically updated using sensor measurement data in a latest predetermined period, a difference between transition of an emotional state using a today's model and transition of an emotional state using a past model, which are calculated by inputting today's sensor measurement data to each of the today's model and the past model, is calculated, a message sentence including advice for improving wellbeing according to the calculated difference is generated, and the generated message sentence is presented to a user.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an information processing device and method, and an information processing system, and is particularly suitable for application to a well-being monitor system that estimates and presents the state of well-being of a user. [Background technology]

[0002] In recent years, methods for achieving well-being have been researched in the field of positive psychology. In addition, a method has been proposed in which a state of well-being is calculated as a quantitative well-being score using a machine learning model that correlates well-being with body acceleration measured by an accelerometer.

[0003] In relation to well-being, for example, Patent Document 1 discloses a mental improvement support device that asks questions or takes measurements of a subject, constructs a model of the subject's actual situation, such as their experiences from the past to the present, based on the answers and results, provides advice packages based on this, and optimizes model construction and advice package identification based on feedback such as answers to questions and satisfaction effects with the advice packages. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2022-63820 Summary of the Invention [Problem to be solved by the invention]

[0005] However, because preferences and stress tolerance, which affect well-being scores, change depending on daily and educational experiences, it is difficult to accurately predict the latest well-being score using a previously constructed model, or to estimate the current state of well-being of a user based on the predicted well-being score.

[0006] Furthermore, even if a well-being score is presented to a user, the user has no idea how to use the well-being score to improve their well-being.

[0007] The present invention has been made in consideration of the above points, and aims to propose an information processing device, method, and information processing system that can accurately estimate a user's state of well-being and present the user with appropriate advice to improve their well-being. [Means for solving the problem]

[0008] In order to solve this problem, in the present invention, an information processing device acquires time-series sensor measurement data output from an acceleration sensor worn by a user and estimates the user's state of well-being based on the acquired sensor measurement data, the information processing device includes a communication device that acquires the sensor measurement data via communication, and a processor that estimates the user's state of well-being based on the sensor measurement data acquired by the communication device, the processor inputs the sensor measurement data into a predetermined model to estimate a transition in emotional states that constitute the state of well-being, periodically updates the model using the sensor measurement data from the most recent predetermined period, inputs today's sensor measurement data into today's model and past models, respectively, calculates the difference between the transition in emotional states using today's model and the transition in emotional states using past models, generates a message text including advice to improve the user's well-being based on the calculated difference, and presents the generated message text to the user.

[0009] Furthermore, in the present invention, in an information processing method executed by an information processing device that acquires time-series sensor measurement data output from an acceleration sensor worn by a user and estimates the user's state of well-being based on the acquired sensor measurement data, the information processing device has a communication device that acquires the sensor measurement data through communication, and a processor that estimates the user's state of well-being based on the sensor measurement data acquired by the communication device, and the information processing method includes a first step in which the processor inputs the sensor measurement data into a predetermined model to estimate a transition in emotional states that constitute the state of well-being, a second step in which the processor periodically updates the model using the sensor measurement data from a recent predetermined period, a third step in which the processor inputs today's sensor measurement data into the model for today and the model for the past, respectively, and calculates the difference between the transition in emotional states using the model for today and the transition in emotional states using the model for the past, and a fourth step in which the processor generates a message text including advice for improving the user's well-being based on the calculated difference, and presents the generated message text to the user.

[0010] Furthermore, in the present invention, an information processing system for estimating a state of well-being of a user includes a terminal that is worn by the user and has an acceleration sensor mounted thereon that measures the acceleration of the user and outputs the measurement results as sensor measurement data, and an information processing device that acquires the sensor measurement data in time series output from the acceleration sensor of the terminal and estimates the state of well-being of the user based on the acquired sensor measurement data, wherein the information processing device includes a communication device that acquires the sensor measurement data through communication, and a processor that estimates the state of well-being of the user based on the sensor measurement data acquired by the communication device. the processor inputs the sensor measurement data into a predetermined model to estimate a transition in the emotional state that constitutes the state of well-being, periodically updates the model using the sensor measurement data from a recent predetermined period, calculates a difference between the transition in the emotional state using the model today and the transition in the emotional state using the model in the past, calculated by inputting today's sensor measurement data into the model for today and the model in the past, generates a message text including advice for improving the well-being according to the calculated difference, and presents the generated message text to the user.

[0011] The information processing device, method, and information processing system of the present invention can determine the user's current state of well-being using a model that is constantly updated according to the user's emotional state at that time, thereby enabling a more accurate estimation of the user's current state of well-being. Furthermore, the information processing device, method, and information processing system of the present invention can easily help the user 6 understand how to improve their lifestyle in the future. [Effects of the Invention]

[0012] According to the present invention, it is possible to realize an information processing device, method, and information processing system that can accurately estimate the state of well-being of a user and present the user with appropriate advice for improving well-being. [Brief explanation of the drawings]

[0013] [Figure 1] 1 is a block diagram showing a configuration of a well-being monitor system according to a first embodiment. [Figure 2] FIG. 1 is a block diagram illustrating an example of the configuration of a wearable terminal. [Figure 3] FIG. 2 is a block diagram showing various databases maintained on a network database server. [Figure 4] FIG. 1 is a block diagram illustrating an example of the configuration of a well-being monitor device. [Figure 5] FIG. 10 is a diagram illustrating a measurement well-being matrix. [Figure 6] 1 is a diagram illustrating an emotional intensity factor, a communication level factor, and an activity level factor. [Figure 7] FIG. 10 is a diagram showing an example of the screen configuration of a one-day review screen. [Figure 8] FIG. 10 is a diagram showing an example of the screen configuration of a well-being tendency display screen. [Figure 9] FIG. 10 is a diagram showing an example of the screen configuration of a model comparison emotion transition screen. [Figure 10] 10 is a chart illustrating the total difference in emotion intensity factor scores between emotion intensity factor score calculation models. [Figure 11] 10 is a diagram illustrating sensor measurement data. [Figure 12] FIG. 10 is a diagram illustrating message generation using message factors. [Figure 13] 10 is a flowchart showing the processing steps for initial emotion calculation model selection processing. [Figure 14] 10 is a flowchart showing the processing procedure of sensor measurement data acquisition and transmission processing. [Figure 15A] 10 is a flowchart showing a processing procedure for one-day review support processing. [Figure 15B] 10 is a flowchart showing a processing procedure for one-day review support processing. [Figure 15C]10 is a flowchart showing a processing procedure for one-day review support processing. [Figure 16A] 10 is a flowchart showing the processing steps of emotion transition screen display processing. [Figure 16B] 10 is a flowchart showing the processing steps of emotion transition screen display processing. [Figure 17] FIG. 10 is a block diagram showing the configuration of a well-being monitor system according to a second embodiment. [Figure 18] FIG. 2 is a block diagram illustrating an example of the configuration of a sensor terminal. [Figure 19] FIG. 2 is a block diagram illustrating an example of the configuration of a mobile terminal. [Figure 20] FIG. 10 is a diagram showing an example of a screen configuration of a well-being tendency display screen according to another embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0014] An embodiment of the present invention will be described in detail below with reference to the drawings.

[0015] (1) First embodiment (1-1) Configuration of Well-Being Monitor System According to This Embodiment In FIG. 1, reference numeral 1 indicates an information processing system according to this embodiment as a whole (hereinafter referred to as a well-being monitor system).

[0016] This well-being monitor system 1 is configured to include a wearable terminal 2, a network database server 3, and a well-being monitor device 4. These wearable terminal 2, network database server 3, and information processing device (hereinafter referred to as the well-being monitor device) 4 are connected to each other via a network 5 such as the Internet, a WAN (Wide Area Network), a LAN (Local Area Network), or a wireless communication line.

[0017] The wearable device 2 is composed of a smart watch or a wearable dedicated wireless communication device that is worn by the user 6 of the well-being monitor system 1. The wearable device 2 has the function of measuring the acceleration in each of the three axial directions (X-axis, Y-axis, and Z-axis) of the user 6 wearing the device, and transmitting the measured acceleration in these three axial directions to the network database server 3 via the network 5.

[0018] In practice, the wearable terminal 2 is configured to include a CPU (Central Processing Unit) 10, a memory 11, an acceleration sensor 12, a touch panel 13, an input / output unit 14, a data wireless transmission / reception unit 15, a network interface unit 16, a clock / calendar unit 17, and a storage device 18, as shown in FIG.

[0019] The CPU 10 is a processor that controls the overall operation of the wearable device 2. The memory 11 is configured, for example, from a volatile semiconductor memory, and is used as a working memory for the CPU 10.

[0020] Acceleration sensor 12 is a sensor that measures the acceleration of user 6 in each of three axial directions and outputs the measurement results as sensor measurement data. Touch panel 13 is an input / output device provided on the surface of wearable terminal 2. Input / output unit 14 is an interface that takes in the sensor measurement data output from acceleration sensor 12 and operation input from user 6 via touch panel 13, and outputs necessary information to touch panel 13 for display.

[0021] The data wireless transceiver 15 is a wireless communication device that complies with long-distance wireless communication standards such as mobile phone wireless communication standards, and communicates with the network database server 3 and well-being monitor device 4 via a relay station such as a base station (not shown) and the network 5. The network interface 16 is an interface that performs various controls including protocol control when transmitting and receiving data via the data wireless transceiver 15, and the clock / calendar unit 17 is a functional unit that counts the current date and time on the wearable device 2.

[0022] The storage device 18 is made of, for example, a non-volatile semiconductor memory or a volatile semiconductor memory, and is used to hold various data and programs that need to be saved. In this embodiment, the storage device 18 stores a sensor measurement data acquisition program 19, a sensor measurement data transmission program 20, and a network browser 21.

[0023] The sensor measurement data acquisition program 19 is a program that causes the CPU 10 to execute a process of acquiring the sensor measurement data output from the acceleration sensor 12 at a predetermined period (for example, a period of 30 Hz, hereinafter referred to as 30 Hz), adding date and time information representing the current date and time to the acquired sensor measurement data, and storing the data in the memory 11.

[0024] In addition, the sensor measurement data transmission program 20 is a program that causes the CPU 10 to execute a process of bundling a predetermined number (tens to hundreds) of sensor measurement data stored in the memory 11 and transmitting them to the network database server 3 via the network interface unit 16 and the data wireless transmission / reception unit 15.

[0025] Furthermore, the network browser 21 is a program that causes the CPU 10 to execute processes such as displaying various screens (described below) provided by the well-being monitor device 4 on the touch panel 13, and transmitting input from the user 6 using the screens displayed on the touch panel 13 to the well-being monitor device 4 via the network interface unit 16 and the data wireless transceiver unit 15.

[0026] On the other hand, the network database server 3 is made up of a general-purpose server device equipped with a CPU, memory, storage device, communication device, etc. As shown in Fig. 3, the storage device of the network database server 3 stores an initial emotion calculation model database 30, a sensor measurement data database 31, an emotion intensity factor score calculation model database 32, a communication level factor score calculation model database 33, a review sentence database 34, and a message factor database 35. Details of these databases will be described later.

[0027] As shown in Figure 4, the well-being monitor device 4 is composed of a general-purpose server device equipped with a CPU 40, memory 41, network interface unit 42, clock / calendar unit 43, input device 44A, display device 44B, input / output unit 45, communication device 46 and storage device 47.

[0028] The CPU 40 is a processor that controls the overall operation of the well-being monitor device 4. The memory 41 is configured, for example, from a volatile semiconductor memory, and is used as a working memory for the CPU 40.

[0029] The network interface unit 42 is configured by, for example, a network interface card (NIC) and performs protocol control during communication with the wearable device 2 and the network database server 3 via the network 5. The clock / calendar unit 43 is a functional unit that counts the current date and time in the well-being monitor device 4.

[0030] The input device 44A is composed of a keyboard, a mouse, etc., and is used by the user 6 to input various operations to the well-being monitor device 4. The display device 44B is composed of a liquid crystal display, an organic EL (Electro-Luminescence) display, etc., and is used to display various information. Note that instead of the input device 44A and the display device 44B, a touch panel that integrates these may be used.

[0031] The input / output unit 45 is an interface having a function of executing signal processing for inputting signals from the input device 44A and signal processing for displaying necessary information on the display device 44B. The communication device 46 is configured by, for example, a network interface card (NIC), and is a communication device having a function of transmitting and receiving data, commands, etc. between the wearable device 2 and the network server 3 via the network 5.

[0032] The storage device 47 is composed of a large-capacity non-volatile storage device such as a hard disk drive or SSD (Solid State Drive), and stores data that needs to be saved, various programs, etc. In the case of this embodiment, the storage device 47 of the well-being monitor device 4 stores an initial emotion calculation model selection program 48, a one-day review processing program 49, and an emotion transition processing program 50 as main programs, and also stores an emotion intensity factor score calculation program 51, a communication level factor score calculation program 52, an activity level factor score calculation program 53, a statistical analysis program 54, a message sentence generation program 55, and a relative analysis program 56 as programs commonly used by these programs. Details of these programs will be described later.

[0033] (1-2) Function of supporting the practice of well-being improvement methods according to this embodiment Next, a description will be given of a well-being improvement technique practice support function installed in the well-being monitor device 4 of this embodiment. In this regard, first, a well-being score that quantifies the state of well-being will be described.

[0034] As shown in Figure 5, the well-being score can be expressed as a point (coordinate) on a table-type matrix (hereinafter referred to as the measurement well-being matrix) with the communication well-being score on the vertical axis and the activity well-being score on the horizontal axis.

[0035] The communication well-being score (WBScore_communication) at a given time is calculated using the following formula:

number

[0036] In addition, the activity well-being score (WBScore_activity) at that time is calculated using the following formula:

number

[0037] Here, the emotion intensity factor is a factor that represents the type of emotion (positive or negative) and its strength at that time, and has a scale of 0 to 1 as shown in Fig. 6. The emotion intensity factor value at a certain time (hereinafter referred to as the emotion intensity factor score) can be calculated by inputting the acceleration parameter value at that time into the emotion intensity factor score calculation model.

[0038] The emotion intensity factor score calculation model here is a model obtained by machine learning the correlation between the acceleration parameter value and the emotion intensity of user 6, based on the acceleration parameters calculated from the time-series sensor measurement data output from the acceleration sensor 12 worn by user 6 and the emotion intensity of user 6 at that time (emotion type and its strength).

[0039] Incidentally, the "acceleration parameters" here include the acceleration intensity, the amount of activity, and the duration of activity and rest.

[0040] "Acceleration strength" is the acceleration strength |a cc |, the acceleration of the user in the X-axis direction at that time measured by the acceleration sensor is a ccx , the acceleration in the Y-axis direction is a ccy , the acceleration in the Z-axis direction is a ccz As a result,

number

[0041] The "amount of activity" is the number of times (the number of zero crossing points) that the acceleration intensity arranged in time series crosses zero per unit time (for example, one minute before that point in time).

[0042] Furthermore, "active / still duration" is the time ("active duration" or "still duration") that an active state (active state) or still state (inactive state) has continued up to that point. "Active / still duration" can be calculated by binarizing the above-mentioned "activity amount" based on its magnitude compared to the average value, with a state higher than the average being an "active state" and a state lower than the average being an "inactive state," with the period during which an "active state" continues being the "active duration" and the period during which an "inactive state" continues being the "still duration."

[0043] The emotion intensity factor score calculation model is a model that calculates acceleration intensity, activity amount, and activity / rest duration each at 30 Hz from time-series sensor measurement data obtained at 30 Hz, and machine-learns the correlation between a total of 16 feature amounts, namely the first quartile, second quartile, third quartile, and interquartile range of the frequency distribution of the calculated acceleration intensity, activity amount, and activity / rest duration over 30 minutes, and the emotion intensity of user 6 at that time. Therefore, by inputting the 30 Hz time-series sensor measurement data into the emotion intensity factor score calculation model, the emotion intensity value of user 6 for every 30 minutes (hereinafter referred to as emotion intensity factor score) is output.

[0044] The communication level factor is a parameter that takes a value of "1" if the user is communicating with others at that time, and "-1" if the user is not. The value of the communication level factor at a certain time (hereinafter referred to as the communication level factor score) can be calculated by inputting the value of the acceleration parameter at that time into the communication level factor score calculation model.

[0045] The communication level factor score calculation model here is a model obtained by machine learning the correlation between each acceleration parameter value and the presence or absence of communication with others, in the same way as the emotion intensity factor score calculation model, using acceleration parameters at a certain point in time and the presence or absence of communication with others as inputs. Therefore, by inputting the 30 Hz time-series sensor measurement data into the communication level factor score calculation model, a value representing the presence or absence of communication with others by user 6 every 30 minutes (hereinafter referred to as the communication level factor score) is output.

[0046] Furthermore, the activity level factor is a parameter that takes a value of "1" if the user 6 is active (expressive) at that time, and "-1" if the user 6 is not active (inner self). The value of the activity level factor (hereinafter referred to as the activity level factor score) can be calculated rule-based (for example, by classifying the standardized amount of activity using a threshold) based on the value of the acceleration parameter at that time. This activity level factor score is also calculated every 30 minutes.

[0047] The well-being improvement method practice support function of this embodiment uses such well-being scores to support the practice of well-being improvement methods, such as the TGT (Three Good Things) exercise, which aim to improve well-being by reflecting on each day.

[0048] The TGT exercise is a well-being improvement method devised by Professor Seligman of the University of Pennsylvania, the founder of positive psychology, and involves listing three good things that happened that day and writing down why they happened before going to bed each night. According to an experiment conducted in 2005, a group that participated in the TGT exercise for one week not only showed an increase in happiness but also a decrease in depression compared to a group that did not participate, and this state was reported to have continued for six months.

[0049] In practice, the well-being monitor device 4 of this embodiment displays a daily review screen 60 such as that shown in FIG. 7 on the touch panel 13 (FIG. 2) of the wearable terminal 2 in response to a predetermined first input operation by the user 6 on the wearable terminal 2 as a method of supporting the practice of the well-being improvement techniques described above.

[0050] This daily review screen 60 is a screen for the user 6 to look back on the changes in their emotions and communication with others throughout the day, and is configured with a measured well-being time chart display area 61, a communication information / emotion information display area 62, a review text input area 63, and an OK button 64.

[0051] The measured well-being time chart display area 61 displays a measured well-being time chart 65 that displays the transition of the emotions of the user 6 throughout the day. In the measured well-being time chart 65, the intensity of the emotions of the user 6 for every 30 minutes estimated by the well-being monitor device 4 based on the sensor measurement data acquired from the wearable device 2 worn by the user 6 (more precisely, the communication well-being score of the user 6 for every 30 minutes calculated by the well-being monitor device 4) is represented by the height of a bar graph 66.

[0052] Furthermore, in the measured well-being time chart 65, bar graphs 66 for each time period during which the well-being monitor device 4 estimates that the user 6 communicated with others based on the sensor measurement data are displayed in a different display format from other time periods. Fig. 7 shows an example in which the bar graphs 66 for each time period during which the well-being monitor device 4 estimates that the user 6 communicated with others are displayed in a darker color than other time periods (for example, 7:30, 9:00 to 11:30, etc.). However, instead of using different shades of color, the bar graphs 66 may also be displayed in different colors.

[0053] On the daily review screen 60, by pressing a position on the measurement well-being time chart 65 corresponding to a desired 30-minute time slot, the user 6 can designate that time slot as the time slot for which the user 6 will look back on that day (hereinafter referred to as the review time slot). At this time, on the measurement well-being time chart 65, an area 65A corresponding to the designated review time slot is colored in a predetermined color, and a character string indicating the time range of the review time slot is displayed in a time slot display area 67 within the communication information / emotion information display area 62.

[0054] Furthermore, the communication information / emotion information display area 62 is provided with a communication information display area 68, and this communication information display area 68 displays the estimation result of the well-being monitor device 4 as to whether or not the user 6 communicated with others during the review time period specified by the user 6 as described above.

[0055] In practice, a first toggle button 69A associated with the case where communication has occurred and a second toggle button 69B associated with the case where communication has not occurred are displayed in the communication information display area 68. The first and second toggle buttons 69A, 69B are a pair of buttons that, when one is pressed to transition the display form to a selected state, transition the other display form to an unselected display form.

[0056] If the well-being monitor device 4 estimates that the user 6 has been communicating with others during the review time period, the first toggle button 69A is displayed in the selected display form, and if the well-being monitor device 4 estimates that the user 6 has not been communicating with others during the review time period, the second toggle button 69B is displayed in the selected display form.

[0057] If the well-being monitor device 4's estimation regarding communication with others is incorrect, the user 6 can correct the information regarding whether or not communication has occurred by pressing the first or second toggle button 69A, 69B corresponding to the correct answer, thereby changing the display form of the first or second toggle button 69A, 69B to the display form when the first or second toggle button 69A, 69B is selected.

[0058] Furthermore, the communication information / emotion information display area 62 also has an emotion information display area 70, and this emotion information display area 70 displays the emotion information of the user 6 during the review time period estimated by the well-being monitor device 4.

[0059] In practice, emotion information display area 70 displays a bar 71A whose left end corresponds to the maximum emotion intensity of a negative emotion and whose right end corresponds to the maximum emotion intensity of a positive emotion, and a slider 71B that can be freely moved from the left end to the right end on bar 71A. Then, in emotion information display area 70, slider 71B is displayed at a position on bar 71A that corresponds to the emotion type (positive or negative) and emotion intensity of user 6 during the review time period, as estimated by well-being monitor device 4.

[0060] If the well-being monitor device 4's estimation of the user 6's emotions during the review time period is incorrect, the user 6 can correct the emotion information by moving the slider 71B to the correct position on the bar 71A. In this case, the height and orientation of the bar graph 66 displayed for the review time period on the measured well-being time chart 65 are automatically corrected according to the correction.

[0061] In this way, the well-being monitor device 4 updates the statistics regarding the well-being estimated for the user 6 based on feedback from the user 6 using the daily review screen 60, thereby making it possible to obtain more accurate information regarding the well-being of the user 6.

[0062] Furthermore, the reflection sentence input area 63 is provided with a third toggle button 72 and a reflection sentence input text box 73. In the reflection sentence input area 63, by pressing the third toggle button 72 to transition its display form to the display form at the time of selection, the user 6 can input a sentence (hereinafter referred to as a reflection sentence) into the reflection sentence input text box 73 explaining the reason why the user 6 felt the emotion of the above-mentioned emotion type and emotion intensity (the emotion of the emotion type and emotion intensity displayed in the emotion information display area) during the reflection time period.

[0063] The user 6 can then correct the estimation results of the well-being monitor device 4 for several time periods as needed in the communication information display area 68 and the emotion information display area 70 as described above, and can also review the day by entering a review sentence in the review sentence input text box 73 in the review sentence input area 63.

[0064] Then, by pressing the OK button 64, the user 6 can input various information about that day, including the reflection text entered on the daily reflection screen 60, as well as information about the emotions and communication progress of the day with others, into the well-being monitor device 4 as reflection information for that day.

[0065] The well-being monitor device 4 determines the well-being trend of the user 6 for that day based on the captured daily review information and the above-mentioned activity well-being score calculated using the sensor measurement data of the user 6 for that day, and displays a well-being trend display screen 80 as shown in Figure 8 on the wearable terminal 2 of the user 6, showing the determination results.

[0066] On this well-being trend display screen 80, the above-mentioned measured well-being matrix 81, which maps the well-being score, which is a combination of the communication well-being score and activity well-being score for every 30 minutes of the day, in heat map format, is displayed on the left side of the screen.

[0067] In the measured well-being matrix 81, a sector-shaped heat map area 82 with a central angle of 90 degrees and centered at the origin is provided in each of the first, second, third, and fourth quadrants.

[0068] The heat map region 82 of each quadrant is divided into multiple regions (arch-shaped regions centered at the origin, hereinafter referred to as arch regions) 82A according to the distance from the center. The greater the number of well-being scores for each 30-minute period mapped to each of these arch regions 82A, the darker the color or density of the color. In other words, in the measurement well-being matrix 81, the frequency distribution of the well-being scores for each 30-minute period mapped to each arch region 82A is represented by color or shade of color.

[0069] Additionally, a neutral frame 83 consisting of a circle of a predetermined size centered at the origin is displayed in a predetermined color in the measured well-being matrix 81. This neutral frame 83 indicates that the well-being of the user 6 is in a neutral state. Therefore, in the measured well-being matrix 81, it is preferable that the well-being score for every 30 minutes is displayed inside the neutral frame 83, and that the heat map region 82 where the communication well-being score and activity well-being score are on the positive side is displayed in a dark color.

[0070] Furthermore, a message field 84 is provided on the right side of the well-being tendency display screen 80, and this message field 84 displays a judgment message from the well-being monitor device 4 regarding the well-being tendency of the user 6 for that day (hereinafter referred to as the well-being tendency judgment message).

[0071] This well-being tendency determination message includes the well-being monitor device 4's determination result of the user 6's well-being tendency for that day based on the well-being score for that day, and advice on what the user 6 should pay attention to in the future in order to improve the well-being of the user 6. Thus, the user 6 can improve their well-being by referring to this well-being tendency determination message and making changes to their own lifestyle in the future as necessary.

[0072] The well-being tendency display screen 80 can be closed by pressing the OK button 85 provided at the bottom right of the screen.

[0073] On the other hand, when the OK button 64 on the daily review screen 60 described above in FIG. 7 is pressed to import the user 6's daily review information, the well-being monitor device 4 updates the emotion intensity factor score calculation model and the communication level factor score calculation model by relearning the user's review information for the past week as learning data.

[0074] Then, the well-being monitor device 4 registers the updated emotion intensity factor score calculation model and communication level factor score calculation model in the network database server 3 (Figure 1) as today's emotion intensity factor score calculation model and communication level factor score calculation model, adding date information of today's date.

[0075] As a result, the emotion intensity factor score calculation model and the communication level factor score calculation model are updated to the latest ones every day, and the updated emotion intensity factor score calculation model and the communication level factor score calculation model are added with date information of the update date and stored in the network database server 3 as the daily emotion intensity factor score calculation model and the communication level factor score calculation model.

[0076] Thereafter, when the user 6 performs a predetermined second input operation, the well-being monitor device 4 reads out the emotion intensity factor score calculation model for today's date and a total of 12 emotion intensity factor score calculation models for the date one month ago, two months ago, ..., and one year ago from the daily emotion intensity factor score calculation models stored in the network database server 3. Note that, hereinafter, each emotion intensity factor score calculation model for the date one month ago, two months ago, ..., and one year ago will be referred to as the "emotion intensity factor score calculation model for each month."

[0077] The well-being monitor device 4 also reads out from the network database server 3 today's sensor measurement data stored in the network database server 3. Then, the well-being monitor device 4 inputs today's sensor measurement data read out from the network database server 3 into the emotion intensity factor score calculation model for each month, and calculates time-series data for today's emotion intensity factor score using the emotion intensity factor score calculation model for each month.

[0078] Then, the well-being monitor device 4 displays a model comparison emotion transition screen 90 as shown in Figure 9 on the touch panel 13 of the wearable device 2 of the user 6 based on the time series data of today's emotion intensity factor score using the emotion intensity factor score calculation model for each month calculated.

[0079] This model comparison emotion transition screen 90 comprises a latest emotion intensity transition display area 91 , an emotion intensity transition display area 92 from one month ago, an emotion intensity transition display area 93 from one year ago, and an annual emotion transition display area 94 .

[0080] The latest emotion intensity transition display area 91 displays an emotion transition graph 100, which shows the transition of emotion intensity over 30 minutes, obtained by inputting today's sensor measurement data into the latest emotion intensity factor score calculation model (i.e., today's emotion intensity factor score calculation model), with the emotion intensity each represented by the height of a bar graph.

[0081] Furthermore, the emotion intensity transition display area 92 one month ago displays an emotion transition graph 101 in which the emotion intensity for each 30 minute interval is represented by the height of each bar graph, obtained by inputting today's sensor measurement data into the emotion intensity factor score calculation model for the date one month prior to today.

[0082] Additionally, the emotion intensity transition display area 92 one month ago is also provided with a message field 102, which displays a message regarding the difference between the well-being one month ago estimated from the transition in emotion intensity every 30 minutes obtained by inputting today's sensor measurement data into the emotion intensity factor score calculation model for the date one month ago, and today's well-being estimated from the transition in emotion intensity today. Note that this message includes the assessment result by the well-being monitor device 4 of the change in the well-being of the user 6 over the past month, and advice for the user 6 to further improve their well-being in the future.

[0083] Furthermore, the emotion intensity transition display area 93 one year ago displays an emotion transition graph 103 in which the emotion intensity for each 30-minute interval is represented by the height of each bar graph, obtained by inputting today's sensor measurement data into the emotion intensity factor score calculation model for the date one year ago from today.

[0084] In addition, the emotion intensity transition display area 93 one year ago also has a message field 104, which displays a message regarding the difference between the well-being one year ago estimated from the transition in emotion intensity every 30 minutes obtained by inputting today's sensor measurement data into the emotion intensity factor score calculation model for the date one year ago, and today's well-being estimated from the transition in emotion intensity today. Note that this message includes the assessment results, made by the well-being monitor device 4, of changes in the well-being of the user 6 over the past year, and advice for the user 6 to further improve their well-being in the future.

[0085] Here, emotion transition graph 101 represents the transition of emotion intensity of user 6 one month ago after a day identical to today has passed, and emotion transition graph 103 represents the transition of emotion intensity of user 6 one year ago after a day identical to today has passed. Therefore, based on these emotion transition graphs 100, 101, and 103, user 6 can compare the transition of emotion intensity today, one month ago, and one year ago when the same event occurs, thereby confirming the effectiveness of the well-being improvement method according to this embodiment. Furthermore, user 6 can also make lifestyle improvements to further enhance well-being based on the messages displayed in each message field 102 and 103.

[0086] Furthermore, the annual emotion transition display area 94 displays a difference total transition graph 105. As shown in Fig. 10, the difference total transition graph 105 is a graph that plots, in order of oldest month, the sum of differences (hereinafter simply referred to as the difference total) between the emotion intensity factor scores for every 30 minutes obtained by inputting today's sensor measurement data into the emotion intensity factor score calculation model for today's date and the emotion intensity factor scores for every 30 minutes for each month obtained by inputting today's sensor measurement data into each emotion intensity factor score calculation model for the date one month ago, two months ago, ..., and one year ago. The difference total transition graph 105 also displays a regression line K1 that represents the linear regression correlation between the plotted difference totals for each month.

[0087] The annual emotion transition display area 94 also has a message field 106, in which a message is displayed showing the diagnosis result of the well-being monitor device 4 for the transition of the user 6's well-being over the course of one year, which is created based on the slope of the regression line K1 and the average (or median) of the above-mentioned monthly difference total values.

[0088] As means for realizing the well-being improvement technique practice support function according to this embodiment, the network database server 3 stores an initial emotion calculation model database 30, a sensor measurement data database 31, an emotion intensity factor score calculation model database 32, a communication level factor score calculation model database 33, a review sentence database 34, and a message factor database 35, as described above with reference to FIG. 3.

[0089] As described above with reference to FIG. 4, the storage device 47 of the well-being monitor device 4 stores the following main programs: an initial emotion calculation model selection program 48, a one-day review processing program 49, and an emotion transition processing program 50. The storage device 47 also stores the following programs commonly used by these programs: an emotion intensity factor score calculation program 51, a communication level factor score calculation program 52, an activity level factor score calculation program 53, a statistical analysis program 54, a message sentence generation program 55, and a relative analysis program 56.

[0090] The initial emotion calculation model database 30 is a database in which a plurality of emotion intensity factor score calculation models and communication level factor score calculation models in an initial state are stored. In the case of this embodiment, in order to use emotion intensity factor score calculation models and communication level factor score calculation models suited to the attributes (including personality) of the user 6 to support the practice of a well-being improvement method, a plurality of types of emotion intensity factor score calculation models and communication level factor score calculation models in an initial state suited to several representative attributes are stored in the initial emotion calculation model database 30.

[0091] In this embodiment, user 6 is first asked to answer a personality questionnaire (a test of attributes including character), and the emotional intensity factor score calculation model and communication level factor score calculation model that are most suitable for the user's attributes determined based on the answers to the personality questionnaire are used as the initial emotional intensity factor score calculation model and communication level factor score calculation model for user 6.

[0092] The sensor measurement data database 31 is a database for storing the sensor measurement data of the acceleration sensor 12 (FIG. 1) transmitted from the wearable device 2 worn by the user 6 as described above. As shown in FIG. 11, the sensor measurement data includes information on the date and time when the sensor measurement data was acquired by the sensor measurement data acquisition program 19 (FIG. 2) of the wearable device 2, as well as the acceleration in the X-axis direction (X-axis acceleration), the acceleration in the Y-axis direction (Y-axis acceleration), and the acceleration in the Z-axis direction (Z-axis acceleration).

[0093] The emotion intensity factor score calculation model database 32 is a database that stores an emotion intensity factor score calculation model for each user 6. The emotion intensity factor score calculation model database 32 stores several months to one year's worth of updated emotion intensity factor score calculation models for each user 6, which are updated daily.

[0094] The communication level factor score calculation model database 33 is a database that stores a communication level factor score calculation model for each user 6. The communication level factor score calculation model database 33 stores several months to one year's worth of updated communication level factor score calculation models for each user 6, which are updated daily.

[0095] The review sentence database 34 is a database for storing sentence data of daily review sentences (hereinafter referred to as review sentence data) entered by the user 6 in the review sentence input text box 73 of the review sentence input area 63 of the daily review screen 60 described above with reference to Fig. 7. The review sentence database 34 stores such daily review sentence data for each user 6.

[0096] The message factor database 35 is a database that stores in advance message factors that are the basis for the well-being tendency judgment message to be displayed in the message field 84 of the well-being tendency display screen 80 described above with reference to Fig. 8, and message factors that are the basis for the messages to be displayed in the message fields 102, 104, and 106 of the model comparison emotion transition screen 90 described above with reference to Fig. 9. In the message factor database 35, message factors that are respectively associated with the values ​​of multiple types of statistics are registered in advance.

[0097] For example, as shown on the right side of Figure 12, for the well-being tendency judgment message displayed in the message field 84 of the well-being tendency display screen 80 described above with reference to Figure 8, the message factors related to the quadrant to which the most well-being scores every 30 minutes are mapped (hereinafter referred to as the maximum frequency quadrant) are stored in the message factor database 35 as follows: for the first quadrant, "actively moving and communicating with others," for the second quadrant, "actively working alone," for the third quadrant, "quietly alone," and for the fourth quadrant, "quietly communicating with others."

[0098] As for message factors relating to the frequency distribution of well-being scores every 30 minutes in the maximum frequency quadrant, for the second quadrant, when the median of the frequency distribution is between 0 and 0.2, the message factors "strong negative feelings, be careful" are associated; when it is between 0.2 and 0.4, the message factors "slightly more negative feelings", when it is between 0.4 and 0.6, the message factors "calm", when it is between 0.6 and 0.8, the message factors "slightly more positive feelings", and when it is between 0.8 and 1, the message factors "positive feelings, but be careful as the emotional utility is too great", and these message factors are stored in message factor database 35.

[0099] Furthermore, as message factors related to the frequency distribution of the well-being score every 30 minutes when the quadrant with the highest frequency is the second quadrant, when the interquartile range of the frequency distribution is between 0 and 0.1, the message factor "this emotion is very strong" is associated with the interquartile range of the frequency distribution between 0.1 and 0.6, the message factor "emotions are moderately spread around this emotion" is associated with the interquartile range of the frequency distribution between 0.1 and 0.6, and the message factor "emotions are widely spread around this emotion" is associated with the interquartile range of the frequency distribution between 0.6 and 1.0, and these message factors are stored in message factor database 35.

[0100] On the other hand, the initial emotion calculation model selection program 48 of the well-being monitor device 4 is a program that causes the CPU 40 to execute a process of selecting an emotion intensity factor score calculation model and a communication level factor score calculation model that are most suitable for the personality of the user 6 from the emotion intensity factor score calculation models and communication level factor score calculation models in the initial state stored in the initial emotion calculation model database 30, based on the answers to the above-mentioned personality questionnaire that the user 6 has taken in advance.

[0101] The daily review processing program 49 is a program that causes the CPU 40 to execute processing for displaying the daily review screen 60 described above with reference to Fig. 7 and the well-being tendency display screen 80 described above with reference to Fig. 8 on the display device 44B (Fig. 4). The emotion transition processing program 50 is a program that causes the CPU 40 to execute processing for displaying the model comparison emotion transition screen 90 described above with reference to Fig. 9 on the display device 44B.

[0102] The emotion intensity factor score calculation program 51 is a program that causes the CPU 40 to execute a process of calculating the emotion intensity factor score of the user 6, for example, every 30 minutes, using an emotion intensity factor score calculation model that is selected for the user 6 by the initial emotion calculation model selection program 48 and updated daily.

[0103] The communication level factor score calculation program 52 is a program that causes the CPU 40 to execute a process of calculating the communication level factor score of the user 6, for example, every 30 minutes, using a communication level factor score calculation model that is selected for the user 6 by the initial emotion calculation model selection program 48 and updated daily.

[0104] Furthermore, the activity level factor score calculation program 53 is a program that causes the CPU 40 to execute a process of calculating the activity level factor score of the user 6, for example, every 30 minutes, in accordance with prescribed rules (for example, using a predetermined activity level factor score calculation formula prepared in advance).

[0105] The statistical analysis program 54 is a program that causes the CPU 40 to execute various statistical analysis operations.

[0106] The message text generation program 55 is a generation AI (Artificial Intelligence) program that causes the CPU 40 to execute processing to generate a well-being tendency judgment message to be displayed in the message field 84 of the well-being tendency display screen 80 described above with reference to FIG. 8, and a message to be displayed in each of the message fields 102, 104, and 106 of the model comparison emotion transition screen 90 described above with reference to FIG. 9.

[0107] For example, as shown in the measured well-being matrix 110 displayed in the upper left of Figure 12, the quadrant with the highest frequency is the second quadrant, the median of the well-being scores for every 30 minutes within that second quadrant (hereinafter referred to as the intra-quadrant median) is 0.7, and the interquartile range of the well-being scores for every 30 minutes within that second quadrant (hereinafter referred to as the intra-quadrant interquartile range) is 0.2.

[0108] In this case, the message text generation program 55 uses the message factor "actively engaged alone" corresponding to the second quadrant among the message factors regarding the maximum frequency quadrant stored in the message factor database 35 of the network database server, the message factor "having a somewhat positive mood" corresponding to 0.6 ≦ f < 0.8 among the message factors regarding the median within the quadrant stored in the message factor database 35, and the message factor "the emotions spreading moderately centered around this emotion" corresponding to 0.1 < f < 0.6 among the message factors regarding the interquartile range within the quadrant stored in the message factor database 35 to create the message displayed at the lower left of FIG. 12.

[0109] Furthermore, the relative analysis program 56 is a program that causes the CPU 40 to execute a process of calculating the total difference value of the emotion intensity factor scores between the daily emotion intensity factor score calculation models described above with respect to FIG. 10.

[0110] (1-3) Various processes executed in relation to the well-being improvement method practice support function Next, the contents of various processes executed in the wearable terminal 2 and the well-being monitor device 4 in relation to such a well-being improvement method practice support function will be described. In the following, the processing entity of each process is a program, but in practice, it is needless to say that the CPU 10 (FIG. 2) of the wearable terminal 2 and the CPU 40 (FIG. 4) of the well-being monitor device 4 execute the process based on the program.

[0111] (1-3-1) Initial emotion calculation model selection process FIG. 13 shows the flow of a series of processes (hereinafter referred to as the initial emotion calculation model selection process) executed by the initial emotion calculation model selection program 48 (FIG. 4) of the well-being monitor device 4 when the user 6 wearing the wearable terminal 2 operates the touch panel 13 (FIG. 2) of the wearable terminal 2 to input an answer to a predetermined personality questionnaire and the answer is transmitted to the well-being monitor device 4.

[0112] When the initial emotion calculation model selection program 48 receives this answer, it commences the initial emotion calculation model selection process shown in Figure 13. Then, the initial emotion calculation model selection program 48 first uses rule-based discrimination processing based on the received answer to select an emotion intensity factor score calculation model and a communication level factor score calculation model that are most suitable for the attributes of that user 6 from among several initial emotion intensity factor score calculation models and communication level factor score calculation models stored in the initial emotion calculation model database 30 (Figure 3) of the network database server 3 (S1).

[0113] Next, the initial emotion calculation model selection program 48 reads out the selected emotion intensity factor score calculation model and communication level factor score calculation model from the initial emotion calculation model database 30 (Fig. 3) of the network database server 3, and stores these read-out emotion intensity factor score calculation models in the emotion intensity factor score calculation model database 32 (Fig. 3) in association with the user 6, and also stores the read-out communication level factor score calculation model in the communication level factor score calculation model database 33 (Fig. 3) in association with the user 6 (S2). Then, the initial emotion calculation model selection program 48 thereafter ends this initial emotion calculation model selection processing.

[0114] (1-3-2) Sensor measurement data acquisition and transmission processing Meanwhile, FIG. 14 shows the flow of a series of processes (hereinafter referred to as sensor measurement data acquisition and transmission processes) executed by the sensor measurement data acquisition program 19 (FIG. 2) in the wearable device 2 worn by the user 6.

[0115] This sensor measurement data acquisition and transmission process starts when the wearable terminal 2, in which the sensor measurement data acquisition program 19 is installed, is worn by the user 6 and powered on.

[0116] Then, the sensor measurement data acquisition program 19 first initializes a transmission counter, which is a variable recorded in the memory 11 (Figure 2) of the wearable terminal 2, and erases all sensor measurement data stored in the memory 11 at that time (S10).

[0117] Next, the sensor measurement data acquisition program 19 waits for a predetermined time (1 / 30 seconds), which is the acquisition cycle of the sensor measurement data (S11).Then, the sensor measurement data acquisition program 19 acquires the sensor measurement data output from the acceleration sensor 12 (FIG. 2) via the input / output unit 14 (FIG. 2), adds information on the current date and time to the acquired sensor measurement data, and stores the data in the memory 11 (S12).

[0118] Next, the sensor measurement data acquisition program 19 increments the count value of the transmission counter by 1 (S13), and determines whether the count value of the transmission counter after the increment is equal to or greater than a preset threshold value (for example, a value of several tens to several hundreds, hereinafter referred to as the collective transmission threshold value) (S14). If the sensor measurement data acquisition program 19 obtains a negative result in this determination, it returns to step S11, and thereafter repeats the processes of steps S11 to S14 until it obtains a positive result in step S14.

[0119] Then, when the number of sensor measurement data stored in memory 11 eventually exceeds the threshold for collective transmission, and the count value of the transmission counter also exceeds the threshold for collective transmission, the sensor measurement data acquisition program 19 obtains a positive result in step S14, and requests the sensor measurement data transmission program 20 (Figure 2) to transmit to the network database server 3 all of the sensor measurement data that has been stored in memory 11 since the last time sensor measurement data was transmitted to the network database server 3 (S15).

[0120] Thus, upon receiving this request, the sensor measurement data transmission program 20 reads all the sensor measurement data to be transmitted from the memory 11, and transmits the read sensor measurement data together to the network database server 3. As a result, the sensor measurement data is stored in the sensor measurement data database 31 (FIG. 3).

[0121] Thereafter, the sensor measurement data acquisition program 19 determines whether or not there has been an input operation to stop the transmission of the sensor measurement data to the network database server 3, for example, by the user 6 performing a predetermined operation on the touch panel 13 (S16).

[0122] If the sensor measurement data acquisition program 19 obtains a negative result in this determination, it returns to step S10 and repeats the processing from step S10 onwards until it obtains a positive result in step S16. If the sensor measurement data acquisition program 19 subsequently obtains a positive result in step S16, it ends this sensor measurement data acquisition and transmission processing.

[0123] (1-3-3) Daily review support processing 15A to 15C show the flow of a series of processes (hereinafter referred to as the daily review support process) executed by the well-being monitor device 4 upon receiving a predetermined request (hereinafter referred to as the daily review screen display request) sent from the wearable device 2 when the user 6 performs a first input operation on the wearable device 2 to display the daily review screen 60 described above with reference to FIG. 7.

[0124] When the well-being monitor device 4 receives this request to display the daily review screen, the daily review support process shown in Figures 15A to 15C is started. Then, first, the daily review process program 49 reads out the latest version of the emotion intensity factor score calculation model (the emotion intensity factor score calculation model updated the previous day) selected for the currently targeted user 6 (hereinafter referred to as the first target user) in the initial emotion calculation model selection process described above with reference to Figure 13 from the emotion intensity factor score calculation model database 32 (Figure 3) of the network database server 3, and stores it in the storage device 47 (S20). Note that if this is the first day of practicing the well-being improvement method according to this embodiment, the process of step S20 is omitted.

[0125] Furthermore, the one-day review processing program 49 reads out the latest version of the communication level factor score calculation model selected for the first target user 6 in the initial emotion calculation model selection processing (the communication level factor score calculation model after updating on the previous day) from the communication level factor score calculation model database 33 (FIG. 3) of the network database server 3, and stores it in the storage device 47 (S21). Note that if it is the first day of practicing the well-being improvement method according to this embodiment, the processing of step S21 is also omitted.

[0126] Next, the one-day review processing program 49 reads all of the time-series sensor measurement data for the first target user 6 for today from the sensor measurement data database 31 (FIG. 3) of the network database server 3 and stores it in the storage device 47 (S22).

[0127] Next, the daily review processing program 49 instructs the emotion intensity factor score calculation program 51 (FIG. 4) to calculate time series data of today's emotion intensity factor scores (S23). Thus, the emotion intensity factor score calculation program 51 calculates time series data of today's emotion intensity factor scores for every 30 minutes by inputting the time series sensor measurement data for the first target user 6 stored in the storage device 47 into the emotion intensity factor score calculation model in the initial state of the first target user 6 saved in the storage device 47 or the emotion intensity factor score calculation model stored in the storage device 47 in step S20.

[0128] Furthermore, the one-day review processing program 49 instructs the communication level factor score calculation program 52 (FIG. 4) to calculate time-series data of the communication level factor scores for today (S24). Thus, the communication level factor score calculation program 52 inputs the time-series sensor measurement data for today of the first target user 6 stored in the storage device 47 in step S22 into the initial state communication level factor score calculation model for the first target user 6 saved in the storage device 47 or the communication level factor score calculation model stored in the storage device 47 in step S21, thereby calculating time-series data of the communication level factor scores for every 30 minutes of today.

[0129] Furthermore, the daily review processing program 49 instructs the activity level factor score calculation program 53 (FIG. 4) to calculate time series data of today's activity level factor scores (S25). Thus, the activity level factor score calculation program 53 calculates time series data of today's activity level factor scores for every 30 minutes of the day by sequentially inputting the time series sensor measurement data for the first target user 6 stored in the storage device 47 in step S22 into a predetermined activity level factor score calculation formula.

[0130] Thereafter, the daily review processing program 49 instructs the statistical analysis program 54 (FIG. 4) to calculate time-series data of today's communication well-being score (S26). Thus, the statistical analysis program 54 uses the time-series data of today's emotion intensity factor score calculated in step S23 and the time-series data of the communication level factor score calculated in step S24 to calculate time-series data of the communication well-being score every 30 minutes using the above-mentioned formula (1). Specifically, the statistical analysis program 54 calculates time-series data of the communication well-being score by multiplying each emotion intensity factor score of today by the communication level factor score of today corresponding to that emotion intensity factor score.

[0131] Next, the daily review processing program 49 instructs the statistical analysis program 54 to calculate time series data of today's activity well-being score (S27). Thus, the statistical analysis program 54 calculates time series data of the activity well-being score every 30 minutes using the time series data of today's emotion intensity factor score calculated in step S23 and the time series data of the activity level factor score calculated in step S25. Specifically, the statistical analysis program 54 calculates the time series data of the activity well-being score by multiplying each emotion intensity factor score of today by the activity level factor score of today corresponding to that emotion intensity factor score.

[0132] Then, the daily review processing program 49 generates screen data for a daily review screen 60 (Figure 7) in which a measured well-being time chart 65 (Figure 7) onto which the time series data of the communication well-being score calculated in step S26 is mapped is displayed in the measured well-being time chart display area 61 (Figure 7), and transmits the generated screen data to the wearable device 2, thereby displaying the daily review screen 60 on the touch panel 13 of the wearable device 2 (S28).

[0133] Thereafter, the daily review processing program 49 determines whether or not a notification given from the wearable device 2 has been received when some operation has been performed in the communication information / emotion information display area 62 on the daily review screen 60 (i.e., whether or not the content displayed in the communication information / emotion information display area 62 has been modified) (S29).

[0134] If the daily review processing program 49 obtains a positive result in this determination, it corrects the time-series data of the communication well-being score of the first target user 6 for today as needed based on the corrections made at that time. The daily review processing program 49 also generates screen data for a daily review screen 60 whose content has been corrected as needed based on the corrections made, and sends the generated screen data to the wearable device 2, thereby displaying the corrected daily review screen 60 on the touch panel 13 of the wearable device 2 (S28).

[0135] On the other hand, if the daily review processing program 49 obtains a negative result in the determination at step S29, it determines whether or not it has received a notification provided from the wearable device 2 when the OK button 64 (FIG. 7) on the daily review screen 60 is pressed (i.e., whether or not the first target user 6 has pressed the OK button 64 on the daily review screen 60) (S30). If the daily review processing program 49 obtains a negative result in this determination, it returns to step S28, and thereafter repeats the processes of steps S28 to S30 until it obtains a positive result in step S30.

[0136] When the first target user 6 eventually presses the OK button 64 on the daily review screen 60, resulting in a positive result in step S30, the daily review processing program 49 reads the sensor measurement data for the first target user 6 for the past week from the sensor measurement data database 31 (Figure 3) of the network database server 3 and stores it in the memory device 47 (S31).

[0137] In addition, the daily review processing program 49 creates an emotion intensity factor score calculation model for today by re-learning (updating) the emotion intensity factor score calculation model for the first target user 6 using the sensor measurement data for the most recent week stored in the storage device 47 in step S31 (S32).

[0138] Furthermore, the daily review processing program 49 creates a communication level factor score calculation model for today by re-learning (updating) the communication level factor score calculation model for the first target user 6 using the sensor measurement data for the most recent week stored in the memory device 47 in step S31 (S33).

[0139] Then, the daily review processing program 49 stores the emotion intensity factor score calculation model for today created in step S32 in the emotion intensity factor score calculation model database 32 (FIG. 3) in the network database server 3, adding date information for today's date (S34).

[0140] The one-day review processing program 49 also stores the communication level factor score calculation model for today created in step S33 in the communication level factor score calculation model database 33 (FIG. 3) in the network database server 3, adding date information for today's date (S35).

[0141] Furthermore, the daily review processing program 49 stores the review text entered by the first target user 6 in the review text input text box 73 (Figure 7) in the review text input area 63 (Figure 7) on the daily review screen 60 displayed on the wearable terminal 2 of the first target user 6 in step S28, along with date information for today's date, in the review text database 34 (Figure 3) in the network database server 3 (S36).

[0142] Next, the daily review processing program 49 instructs the statistical analysis program 54 to execute a first statistical analysis process to display the measurement well-being matrix 81 on the well-being tendency display screen 80 described above with reference to FIG. 8 (S37).

[0143] Thus, the statistical analysis program 54 uses the time-series data of the communication well-being score calculated in step S26 and the time-series data of the activity well-being score calculated in step S27 to calculate the frequency distribution of the well-being score for each 30-minute period in each quadrant of the measured well-being matrix 81 (FIG. 8). Specifically, the statistical analysis program 54 counts the number of well-being scores for each 30-minute period that are mapped to each arch region 82A (FIG. 8) in each quadrant.

[0144] Furthermore, the daily review processing program 49 instructs the statistical analysis program 54 to execute a second statistical analysis process for generating a well-being tendency judgment message to be displayed in the message field 84 of the well-being tendency display screen 80 (S38).

[0145] Thus, based on the time-series data of the communication well-being score calculated in step S26 and the time-series data of the activity well-being score calculated in step S27, the statistical analysis program 54 detects the quadrant (maximum frequency quadrant) to which the most well-being scores for each 30 minutes are mapped among the quadrants of the measured well-being matrix 81 (FIG. 8).The statistical analysis program 54 also calculates the median of the well-being scores mapped to the maximum frequency quadrant and the interquartile range of the well-being scores in the maximum frequency quadrant.

[0146] Next, the daily review processing program 49 selects a message factor for generating a well-being trend judgment message to be displayed in the message field 84 of the well-being trend display screen 80 based on the processing results of the second statistical analysis processing described above executed by the statistical analysis program 54 (S39).

[0147] Specifically, the daily review processing program 49 selects a message factor associated with the quadrant with the highest frequency of the well-being score for each 30 minutes detected by the statistical analysis program 54 from the message factors associated with each quadrant displayed in the upper right section of Fig. 12, for example, which are stored in the message factor database 35 of the network database server 3. In the example of Fig. 12, the second quadrant is identified as the quadrant with the highest frequency, and the message factor "actively doing activities alone" associated with the second quadrant is selected.

[0148] Furthermore, the daily review processing program 49 selects a message factor associated with the arch region 82A (FIG. 8) containing the actual median of the well-being score calculated by the statistical analysis program 54 from among various message factors related to the median of the well-being score, such as those displayed in the middle right section of FIG. 12, which are stored in the message factor database 35 of the network database server 3. In the example of FIG. 12, the median is 0.7, so the message factor "I feel a little more positive" associated with the arch region 82A in the second quadrant where 0.6≦f<0.8 is selected.

[0149] Furthermore, the daily review processing program 49 selects a message factor corresponding to the interquartile range of the actual well-being score calculated by the statistical analysis program 54 from various message factors related to the interquartile range of the well-being score, such as those displayed in the lower right section of Fig. 12, which are stored in the message factor database 35 of the network database server 3. In the example of Fig. 12, the interquartile range in the maximum frequency quadrant is 0.2, so the message factor "Emotions are spread moderately around this emotion" which corresponds to a range greater than 0.1 and less than 0.6 is selected.

[0150] Thereafter, the daily review processing program 49 instructs the message text generation program 55 (FIG. 4) to generate a well-being tendency judgment message to be displayed in the message field 84 of the well-being tendency display screen 80 based on each message factor selected as described above (S40). Thus, the message text generation program 55 generates a well-being tendency judgment message such as that shown in the lower left of FIG. 12 using generation AI technology based on each of these message factors.

[0151] Next, the daily review processing program 49 generates screen data for the well-being tendency display screen 80 based on the processing result of the statistical analysis program 54 in step S37 and the well-being tendency determination message generated by the message text generation program 55 in step S40. The daily review processing program 49 also transmits the generated screen data to the wearable device 2 of the first target user 6, thereby displaying the well-being tendency display screen 80 based on the image data on the touch panel 13 of the wearable device 2 (S41).

[0152] Thereafter, the daily review processing program 49 waits for the OK button 85 (FIG. 8) on the well-being tendency display screen 80 to be pressed (S42). When the OK button 85 on the well-being tendency display screen 80 is pressed, the daily review processing program 49 ends the daily review support processing.

[0153] (1-3-4) Model comparison emotion transition screen display processing 16A and 16B show the flow of a series of processes (hereinafter referred to as model comparison emotion transition screen display processing) executed by the emotion transition processing program 50 (FIG. 4) of the well-being monitor device 4 that receives a predetermined request (hereinafter referred to as model comparison emotion transition screen display request) sent from the wearable device 2 when a user (hereinafter referred to as the second target user) 6 performs a second input operation on the wearable device 2 to display the model comparison emotion transition screen 90 described above with reference to FIG. 9.

[0154] When the emotion transition processing program 50 recognizes that the well-being monitor device 4 has received the request to display the model comparison emotion transition screen, it starts the model comparison emotion transition screen display processing shown in FIGS. 16A and 16B.

[0155] The emotion transition processing program 50 first reads the emotion intensity calculation models for today and each month up to one year ago for the second target user 6 from the emotion intensity factor score calculation model database 32 (FIG. 3) in the network database server 3 (S50). The emotion transition processing program 50 also reads the sensor measurement data for today for the second target user 6 from the sensor measurement data database 31 (FIG. 3) in the network database server 3 (S51).

[0156] Next, the emotion transition processing program 50 instructs the emotion intensity factor score calculation program 51 (FIG. 4) to calculate time-series data of emotion intensity factor scores using the sensor measurement data for today read in step S51 (S52). Thus, the emotion intensity factor score calculation program 51 inputs the sensor measurement data read by the emotion transition processing program 50 in step S51 into the emotion intensity factor score calculation models for today and for each month up to one year ago read by the emotion transition processing program 50 in step S50, and calculates time-series data of emotion intensity factor scores when using the emotion intensity factor score calculation models for today and for each month up to one year ago.

[0157] Next, the emotion transition processing program 50 instructs the relative analysis program 56 (Fig. 4) to calculate the monthly total difference value as described above with reference to Fig. 10 (S53). Thus, the relative analysis program 56 calculates, for each month, the difference in hourly values ​​between the time-series data of today's emotion intensity factor score calculated using today's emotion intensity factor score calculation model in step S52 and the time-series data of emotion intensity factor score calculated using the emotion intensity factor score calculation model for each month from today up to one year ago in step S52, and calculates, for each month, a total difference value by adding up the calculated hourly differences (S53).

[0158] The emotion transition processing program 50 also instructs the statistical analysis program 54 (FIG. 4) to calculate message statistics for data from one month ago (S54). Thus, the statistical analysis program 54 calculates the absolute value and sign of the difference total value from one month ago calculated by the relative analysis program 56 in step S53 as the message statistics.

[0159] Then, based on the message statistics (absolute values ​​and signs of the difference totals) for the data from one month ago calculated by the statistical analysis program 54 in step S54, the emotion transition processing program 50 selects message factors associated with the values ​​of these message statistics from the message factors stored in the message factor database 35 of the network database server 3 (S55).

[0160] Furthermore, emotion transition processing program 50 instructs message text generation program 55 to generate a message (S56). Thus, using all message factors selected by emotion transition processing program 50 in step S55, message text generation program 55 generates a message for the difference between the time series data of emotion intensity factor scores calculated using the emotion intensity factor score calculation model for today and the time series data of emotion intensity factor scores calculated using the emotion intensity factor score calculation model for one month ago.

[0161] Next, the emotion transition processing program 50 instructs the statistical analysis program 54 (FIG. 4) to calculate message statistics for the data from one year ago (S57). Thus, the statistical analysis program 54 calculates the absolute value and sign of the difference total value from one year ago calculated by the relative analysis program 56 in step S53 as the message statistics.

[0162] Then, based on the message statistics (absolute values ​​and signs of the difference totals) for the data from one year ago calculated by the statistical analysis program 54 in step S57, the emotion transition processing program 50 selects message factors associated with the values ​​of these message statistics from the message factors stored in the message factor database 35 of the network database server 3 (S58).

[0163] Furthermore, emotion transition processing program 50 instructs message text generation program 55 to generate a message (S59). Thus, using all message factors selected by emotion transition processing program 50 in step S55, message text generation program 55 generates a message for the difference between the time series data of emotion intensity factor scores calculated using the emotion intensity factor score calculation model for today and the time series data of emotion intensity factor scores calculated using the emotion intensity factor score calculation model for one year ago.

[0164] Next, the emotion transition processing program 50 instructs the statistical analysis program 54 to calculate correlation statistics of the monthly difference totals calculated in step S53 (S60). Thus, the statistical analysis program 54 calculates correlation statistics between the monthly difference totals calculated in step S53. Specifically, the statistical analysis program 54 calculates the slope of the regression line K1 described above with reference to Figure 9, the average or median of the monthly difference totals, and the sign of the average or median value as these correlation statistics.

[0165] Furthermore, the emotion transition processing program 50 selects, from the message factors stored in the message factor database 35 of the network database server 3, message factors corresponding to each correlation statistic calculated in step S60 (the slope of the regression line K1, the average or median of the difference totals for each month, and the sign of the average or median) (S61).

[0166] Then, the emotion transition processing program 50 instructs the message text generation program 55 to generate a message for the transition of well-being using each message factor selected in step S61 (S62). Thus, the message text generation program 55 generates a message using each message factor selected by the emotion transition processing program 50 in step S62.

[0167] Thereafter, the emotion transition processing program 50 generates screen data for the model comparison emotion transition screen 90 described above with reference to FIG. 9 based on the information obtained as described above, and transmits the generated screen data to the corresponding wearable device 2, thereby displaying the model comparison emotion transition screen 90 on the touch panel 13 (FIG. 2) of that wearable device 2 (S63).

[0168] Specifically, in step S52, the emotion transition processing program 50 generates an emotion transition graph 100 using time-series data of emotion intensity factor scores obtained by inputting today's sensor measurement data into today's emotion intensity factor score calculation model, and places the generated emotion transition graph 100 in the latest emotion intensity transition display area 91.

[0169] Furthermore, in step S52, the emotion transition processing program 50 generates an emotion transition graph 101 using time-series data of emotion intensity factor scores obtained by inputting today's sensor measurement data into the emotion intensity factor score calculation model one month ago, and places the generated emotion transition graph 101 in the emotion intensity transition one month ago display area 92. Furthermore, in step S56, the emotion transition processing program 50 displays the message generated by the message text generation program 55 in the message field 104.

[0170] Furthermore, in step S52, the emotion transition processing program 50 generates an emotion transition graph 103 using time-series data of emotion intensity factor scores obtained by inputting today's sensor measurement data into the emotion intensity factor score calculation model for one year ago, and places the generated emotion transition graph 103 in the emotion intensity transition one year ago display area 93. Furthermore, in step S59, the emotion transition processing program 50 displays the message generated by the message text generation program 55 in the message field 104.

[0171] Furthermore, in step S53, the emotion transition processing program 50 generates a difference total transition graph 105 based on the monthly difference total value calculated by the relative analysis program 56, and places the generated difference total transition graph 105 in the annual emotion transition display area 94. In step S63, the emotion transition processing program 50 also displays the message generated by the message sentence generation program 55 in the message field 106.

[0172] The emotion transition processing program 50 then generates screen data for the model comparison model comparison emotion transition screen created as described above, and transmits the generated screen data to the corresponding wearable device 2, thereby displaying the model comparison model comparison emotion transition screen on the touch panel 13 (Figure 2) of that wearable device 2.

[0173] Then, the emotion transition processing program 50 thereafter ends this emotion transition screen display processing.

[0174] (1-4) Effects of this embodiment As described above, the well-being monitor system 1 of this embodiment updates the emotion intensity factor score calculation model and the communication level factor score calculation model daily based on the measurement data for the most recent week. The well-being monitor system 1 also calculates the difference between the transition in the emotional state of the user 6 using today's emotion intensity factor score calculation model, which is calculated by inputting today's sensor measurement data into today's emotion intensity factor score calculation model and each emotion intensity factor score calculation model from one month ago and one year ago, and the transition in the emotional state of the user 6 using each emotion intensity factor score calculation model from one month ago and one year ago, and displays a message of advice for improving well-being according to the calculated difference.

[0175] Therefore, according to the present well-being monitor system 1, the current well-being state of the user 6 can be determined using an emotion intensity factor score calculation model and a communication level factor score calculation model that are always updated according to the emotional state of the user 6 at that time, and therefore the well-being state of the user 6 can be estimated more accurately.

[0176] Furthermore, according to this well-being monitor system 1, a message text (message text) containing advice for improving well-being is also displayed, allowing the user 6 to easily understand how to improve their life in the future.

[0177] Therefore, according to the present well-being monitor system 1, it is possible to accurately estimate the state of well-being of the user and present the user with appropriate advice for improving their well-being.

[0178] (2) Second embodiment 17, in which the same reference numerals are assigned to parts corresponding to those in Fig. 1, shows a well-being monitor system 120 according to the second embodiment. This well-being monitor system 120 differs from the well-being monitor system 1 of the first embodiment in that a sensor terminal 121 and a mobile terminal 122 are provided instead of the wearable terminal 2.

[0179] The sensor terminal 121 is a wearable terminal such as a smart watch that is worn by a user of the well-being monitor system 120, and as shown in FIG. 18, is configured with a CPU 130, memory 131, an acceleration sensor 132, an input / output unit 133, a data wireless transmission / reception unit 134, a network interface unit 135, a clock / calendar unit 136, and a storage device 137.

[0180] The CPU is a processor that controls the overall operation of the sensor terminal 121. The memory 131 is configured, for example, from a volatile semiconductor memory, and is used as a working memory for the CPU .

[0181] The acceleration sensor 132 is a sensor that measures the acceleration of the user 6 in each of the three axial directions and outputs the measurement results as sensor measurement data. The input / output unit 133 is an interface that has a function of importing the sensor measurement data output from the acceleration sensor 132.

[0182] The data wireless transceiver 134 is a wireless communication device that performs short-range wireless communication, such as Bluetooth, with the mobile terminal 122. The network interface 135 is an interface that performs various controls when transmitting data to the mobile terminal 122 via the data wireless transceiver 134, and the clock / calendar unit 136 is a functional unit that counts the current date and time in the sensor terminal 121.

[0183] The storage device 137 is configured, for example, from a non-volatile semiconductor memory and a volatile semiconductor memory, and is used to hold data that needs to be saved and various programs. In this embodiment, the storage device 137 stores a sensor measurement data acquisition program 138 and a sensor measurement data transmission program 139.

[0184] The sensor measurement data acquisition program 138 and the sensor measurement data transmission program 139 are programs having the same functions as the sensor measurement data acquisition program 19 and the sensor measurement data transmission program 20 of the first embodiment described above with reference to FIG. 2, respectively.

[0185] Thus, similar to the wearable terminal 2 of the first embodiment, the CPU 130 acquires the sensor measurement data output from the acceleration sensor 132 at a predetermined period (30 Hz period) based on the sensor measurement data acquisition program 138, adds date and time information representing the current date and time to the acquired sensor measurement data, and stores it in the memory 131.

[0186] Furthermore, based on the sensor measurement data transmission program 139, the CPU 130 transmits a predetermined number (tens to hundreds) of sensor measurement data stored in the memory 131 to the mobile terminal 122 via the network interface unit 135 and the data wireless transmission / reception unit 134.

[0187] The mobile terminal 122 is configured from a portable communication terminal device such as a smartphone, tablet, etc. As shown in Fig. 19 , the mobile terminal 122 is configured to include a CPU 140, a memory 141, a touch panel 142, an input / output unit 143, a sensor data wireless transmission / reception unit 144, a GUI data wireless transmission / reception unit 145, a network interface unit 146, and a storage device 147.

[0188] The CPU 140 is a processor that controls the overall operation of the mobile terminal 122. The memory 141 is configured, for example, from a volatile semiconductor memory, and is used as a working memory for the CPU 140.

[0189] The touch panel 142 is an input / output device provided on the surface of the mobile terminal 122. The input / output unit 143 is an interface having a function of taking in operation input from the user 6 via the touch panel 142 and outputting necessary information to the touch panel 142 for display.

[0190] The sensor data wireless transceiver 144 is a wireless communication device that performs short-range wireless communication, such as Bluetooth, with the sensor terminal 121. The GUI data wireless transceiver 145 is a wireless communication device that complies with the mobile phone wireless communication standard and has the function of communicating with the network database server 3 and well-being monitor device 4 via a base station (not shown) and the network 5. The network interface 146 is an interface that performs various controls when transmitting and receiving data via the sensor data wireless transceiver 144 and the GUI data wireless transceiver 145.

[0191] The storage device 147 is configured, for example, from a non-volatile semiconductor memory and a volatile semiconductor memory, and is used to hold data that needs to be saved and various programs. In this embodiment, the storage device 147 stores a sensor measurement data receiving program 148, a sensor flag sending program 149, and a network browser 150.

[0192] The sensor measurement data receiving program 148 is a program that causes the CPU 140 to execute a process for receiving sensor measurement data transmitted from the sensor terminal 121. The sensor flag transmitting program 149 is a program that causes the CPU 140 to execute a process for transferring the received sensor measurement data to the network database server 3.

[0193] Furthermore, the network browser 150 is a program having the function of displaying on the touch panel 142 the one-day review screen 60 described above with reference to FIG. 7, the well-being trend display screen 80 described above with reference to FIG. 8, and the model comparison model comparison emotion transition screen described above with reference to FIG. 9, which are provided by the well-being monitor device 4.

[0194] In the well-being monitor system 120 of this embodiment having the above configuration, sensor measurement data similar to the sensor measurement data according to the first embodiment acquired by the sensor terminal 121 is transferred to the network database server 3 via the mobile terminal 122. Therefore, the well-being monitor system 120 of this embodiment can also achieve the same effects as the well-being monitor system 1 of the first embodiment.

[0195] (3) Other embodiments In the first and second embodiments described above, the well-being monitor device 4 according to this embodiment is described as being configured using a single computer device, but the present invention is not limited to this, and the well-being monitor device 4 may also be constructed using a distributed computing system made up of multiple computer devices.

[0196] 8, the well-being trend for the day is determined based on the sensor measurement data for that day, and a well-being trend display screen 80 showing the determination results is displayed on the wearable device 2 of the user 6 as shown in Fig. 8. However, the present invention is not limited to this. For example, the well-being trend for a certain period, such as one week, may be determined based on the sensor measurement data for that period, and the results may be displayed in a measurement well-being matrix 160 similar to the measurement well-being matrix 81 of the first embodiment as shown in Fig. 20, or a well-being trend display screen 162 showing a message from the well-being monitor device 4 in response to the results in a message field 161 may be displayed on the touch panel 13, 142 by a predetermined operation input.

[0197] Furthermore, in the first and second embodiments described above, the network database server 3 and the well-being monitor device 4 are described as being provided separately, but the present invention is not limited to this, and the network database server 3 and the well-being monitor device 4 may also be integrated.

[0198] Furthermore, in the above-described first and second embodiments, the emotion intensity factor score calculation model and the communication level factor score calculation model are updated daily based on the most recent one week's worth of data measurement data. However, the present invention is not limited to this, and the emotion intensity factor score calculation model and the communication level factor score calculation model may be updated based on data measurement data for a period other than one week.

[0199] Furthermore, in the first and second embodiments described above, the communication well-being scores of user 6 for every 30 minutes are displayed on the daily review screen 60 described above with reference to FIG. 7 and the model comparison emotion transition screen 90 described above with reference to FIG. 9, and a measured well-being matrix 81 in which the well-being scores for every 30 minutes are mapped is displayed on the well-being tendency display screen 80 described above with reference to FIG. 8. However, the present invention is not limited to this, and it is also possible to display communication well-being scores for periods other than 30 minutes, or to display a measured well-being matrix 81 in which the well-being scores for periods other than 30 minutes are mapped.

[0200] Furthermore, in the above-described first and second embodiments, the case has been described in which the emotion intensity factor score calculation model and the communication level factor score calculation model are updated in response to the well-being monitor device 4 receiving a request to display a one-day review screen as a trigger. However, the present invention is not limited to this, and for example, the emotion intensity factor score calculation model and the communication level factor score calculation model may be updated at a fixed time during the day. In short, as long as the emotion intensity factor score calculation model and the communication level factor score calculation model are updated (to a certain extent) regularly or (to a certain extent) irregularly, a wide variety of timings can be applied as the timing for updating the emotion intensity factor score calculation model and the communication level factor score calculation model. [Industrial Applicability]

[0201] The present invention can be widely applied to well-being monitor systems of various configurations that estimate and present the state of well-being of a user. [Explanation of symbols]

[0202] 1,120...Well-being monitor system, 2...Wearable terminal, 3...Network database server, 4...Well-being monitor device, 6...User, 10,40...CPU, 13,142...Touch panel, 30...Initial emotion calculation model database, 31...Sensor measurement data database, 32...Emotion intensity factor score calculation model database, 33...Communication level factor score calculation model database, 34...Review sentence database, 35...Message factor database, 48...Initial emotion calculation model selection program RAM, 49...Daily review processing program, 50...Emotion transition processing program, 51...Emotion intensity factor score calculation program, 52...Communication level factor score calculation program, 53...Activity level factor score calculation program, 54...Statistical analysis program, 55...Message text generation program, 56...Relative analysis program, 60...Daily review screen, 80,162...Well-being trend display screen, 81,160...Measured well-being matrix, 90...Model comparison emotion transition screen, 121...Sensor terminal, 122...Mobile terminal.

Claims

1. An information processing device that acquires time-series sensor measurement data output from an acceleration sensor worn by a user and estimates a state of well-being of the user based on the acquired sensor measurement data, a communication device that acquires the sensor measurement data through communication; a processor that estimates a state of well-being of the user based on the sensor measurement data acquired by the communication device; Equipped with The processor: inputting the sensor measurement data into a predetermined model to estimate a transition of an emotional state constituting the state of well-being; periodically updating the model using the sensor measurement data from a recent predetermined period; calculating a difference between a transition of the emotional state using the model for today and a transition of the emotional state using the model for the past, the transition being calculated by inputting the sensor measurement data for today into the model for today and the model for the past; A message text including advice for improving the well-being according to the calculated difference is generated, and the generated message text is presented to the user.

1. An information processing device comprising:

2. The model is The model is obtained by machine learning the correlation between the acceleration parameter value calculated from the time-series sensor measurement data and the emotional state of the user.

2. The information processing apparatus according to claim 1, wherein:

3. The processor: The model is updated based on feedback from the user regarding the transition of the estimated emotional state of the user.

2. The information processing apparatus according to claim 1, wherein:

4. a message factor database is provided in which message factors corresponding to the values ​​of a plurality of types of statistics relating to the state of well-being are registered in advance; The processor: The message factor corresponding to the difference between the transition of the emotional state using the model today and the transition of the emotional state using the model in the past is read from the message factor database, and the message text is generated based on the read message factor.

2. The information processing apparatus according to claim 1, wherein:

5. The processor: inputting the sensor measurement data of today into each of a plurality of past models, and calculating a transition of the emotional state using the model of today and a transition of the emotional state using each of a plurality of past models, Calculating a difference between the transition of the emotional state calculated using the model for today and the transition of the emotional state calculated using each of the past models for each of the past models; The difference between the transition of the emotional state calculated using the model for today and the transition of the emotional state calculated using each of the past models is plotted in order of oldest or newest model used, and the graph is presented together with the message text.

2. The information processing apparatus according to claim 1, wherein:

6. a message factor database is provided in which message factors corresponding to the values ​​of a plurality of types of statistics relating to the state of well-being are registered in advance; The processor: The message factors corresponding to the slope of the regression line of each of the differences plotted on the graph and the absolute value and sign of each of the differences are read from the message factor database, and the message text is generated based on the read message factors.

6. The information processing apparatus according to claim 5,

7. An information processing method executed by an information processing device that acquires time-series sensor measurement data output from an acceleration sensor worn by a user and estimates a well-being state of the user based on the acquired sensor measurement data, The information processing device includes: a communication device that acquires the sensor measurement data through communication; a processor that estimates a state of well-being of the user based on the sensor measurement data acquired by the communication device; and a first step in which the processor inputs the sensor measurement data into a predetermined model to estimate a transition of the emotional state that constitutes the state of well-being; a second step in which the processor periodically updates the model using the sensor measurement data from a recent predetermined period; a third step in which the processor calculates a difference between a transition of the emotional state using the model today and a transition of the emotional state using the model in the past, the transition being calculated by inputting the sensor measurement data of the day into the model today and the model in the past; a fourth step in which the processor generates a message text including advice for improving the well-being according to the calculated difference, and presents the generated message text to the user; An information processing method comprising:

8. The model is The model is obtained by machine learning the correlation between the acceleration parameter value calculated from the time-series sensor measurement data and the emotional state of the user.

8. The information processing method according to claim 7,

9. In the second step, the processor: The model is updated based on feedback from the user regarding the transition of the estimated emotional state of the user.

8. The information processing method according to claim 7,

10. The information processing device includes: a message factor database in which message factors respectively associated with values ​​of a plurality of types of statistics relating to the state of well-being are registered in advance; In the fourth step, the processor: The message factor corresponding to the difference between the transition of the emotional state using the model today and the transition of the emotional state using the model in the past is read from the message factor database, and the message text is generated based on the read message factor.

8. The information processing method according to claim 7,

11. In the first step, the processor inputting the sensor measurement data of today into each of a plurality of past models, and calculating a transition of the emotional state using the model of today and a transition of the emotional state using each of a plurality of past models, In the third step, the processor: Calculating a difference between the transition of the emotional state calculated using the model for today and the transition of the emotional state calculated using each of the past models for each of the past models; In the fourth step, the processor: The difference between the transition of the emotional state calculated using the model for today and the transition of the emotional state calculated using each of the past models is plotted in order of oldest or newest model used, and the graph is presented together with the message text.

8. The information processing method according to claim 7,

12. The information processing device includes: a message factor database in which message factors respectively associated with values ​​of a plurality of types of statistics relating to the state of well-being are registered in advance; In the fourth step, the processor: The message factors corresponding to the slope of the regression line of each of the differences plotted on the graph and the absolute value and sign of each of the differences are read from the message factor database, and the message text is generated based on the read message factors.

12. The information processing method according to claim 11.

13. An information processing system for estimating a user's well-being state, a terminal equipped with an acceleration sensor that is worn by the user, measures the acceleration of the user, and outputs the measurement result as sensor measurement data; an information processing device that acquires the time-series sensor measurement data output from the acceleration sensor of the terminal and estimates a well-being state of the user based on the acquired sensor measurement data; and The information processing device includes: a communication device that acquires the sensor measurement data through communication; a processor that estimates a state of well-being of the user based on the sensor measurement data acquired by the communication device; Equipped with The processor: inputting the sensor measurement data into a predetermined model to estimate a transition of an emotional state constituting the state of well-being; periodically updating the model using the sensor measurement data from a recent predetermined period; calculating a difference between a transition of the emotional state using the model for today and a transition of the emotional state using the model for the past, the transition being calculated by inputting the sensor measurement data for today into the model for today and the model for the past; A message text including advice for improving the well-being according to the calculated difference is generated, and the generated message text is presented to the user. An information processing system comprising:

Images