Well-being score generation system and method

Abstract

To properly digitize various personal well-being.SOLUTION: A well-being score generation system is configured to: store in advance, multiple well-being score calculation models which are models for calculating a well-being score with a feature amount related to activity of the body of a user as input, the models being estimation models corresponding to an activity factor group indicating a person type expressed in a mutual relationship between activity briskness, and positive emotion or interpersonal communication; acquire a score indicating characteristics of personality of the user, on the basis of user solution information to the personality test; identify an activity factor group to which the user corresponds on the basis of the score; acquire sensor data of the user; calculate the feature amount of the user on the basis of the sensor data; select a well-being score calculation model on the basis of the activity factor group to which the user corresponds; and calculate the well-being score by inputting the feature amount of the user as input, to the selected well-being score calculation model.SELECTED DRAWING: Figure 6

Description

[Technical field]

[0001] The present disclosure relates to a technique for quantifying a degree of happiness. [Background technology]

[0002] In the field of positive psychology, methods for realizing well-being have been studied. However, since well-being is related to human emotions, it is not easy to quantitatively measure the level of well-being of a person or how much the well-being of a person has improved. Patent Documents 1 and 2 disclose methods for estimating a user's emotions from data acquired by various sensors. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2018-138155 A [Patent Document 2] JP 2018-139087 A Summary of the Invention [Problem to be solved by the invention]

[0004] In recent years, there is a demand for technology that quantitatively measures the well-being of individuals in daily life, not just for use in organizations such as companies. However, the methods disclosed in Patent Document 1 and Patent Document 2 are primarily intended for use in organizations, and are methods for calculating team emotions, which are the emotional atmosphere of a place fostered by collaborative activities among multiple people. These methods are not suitable for estimating the emotions of individuals, including those in daily life. Although there are technologies for estimating the emotions of individuals, there are no methods that can appropriately quantify the well-being of diverse people with different emotional characteristics.

[0005] One objective of the present disclosure is to provide a technique for appropriately quantifying the well-being of a variety of individuals. [Means for solving the problem]

[0006] A well-being index score creation system according to one aspect included in the present disclosure is a well-being score generation system that generates a well-being score that numerically represents the intensity of a user's well-being from sensor data measuring the user's condition, the well-being score generation system having a memory device and a processing device, the memory device is a well-being score calculation model that calculates a well-being score using features related to the user's physical activity as input, and stores a plurality of well-being score calculation models that are estimation models corresponding to activity factor groups that indicate a person type that is expressed in the mutual relationship between activity activity and positive emotions or interpersonal communicability, the processing device acquires user answer information in which the user's answers to a personality test are recorded, acquires a personality factor score that is a score that represents the characteristics of the user's personality based on the user answer information, identifies the activity factor group that the user corresponds to based on the personality factor score, acquires sensor data of the user, calculates the user's features based on the sensor data, selects the well-being score calculation model based on the activity factor group that the user corresponds to, and calculates the well-being score by inputting the user's features into the selected well-being score calculation model.

[0007] A well-being index score creation method according to one aspect included in the present disclosure is a well-being score generation method for generating a well-being score that numerically represents the intensity of a user's well-being from sensor data measuring a user's condition using a system having at least a storage device and a processing device, in which a computer calculates a well-being score using features related to the user's physical activity as input, and stores a plurality of well-being score calculation models that are estimation models corresponding to activity factor groups that indicate a person type that is expressed in the mutual relationship between activity activity and positive emotions or interpersonal communicability, acquires user answer information in which the user's answers to a personality test are recorded, and acquires a personality factor score that is a score that represents the characteristics of the user's personality based on the user answer information, identifies the activity factor group that the user corresponds to based on the personality factor score, acquires sensor data of the user, calculates the features of the user based on the sensor data, selects the well-being score calculation model based on the activity factor group that the user corresponds to, and calculates the well-being score using the features of the user as input to the selected well-being score calculation model. Effect of the Invention

[0008] According to one aspect of the present disclosure, well-being can be appropriately quantified. [Brief description of the drawings]

[0009] [Figure 1] FIG. 1 is a conceptual diagram illustrating a configuration example of a well-being score generation system according to an embodiment of the present disclosure. [Diagram 2] FIG. 2 is a block diagram illustrating a configuration example of a terminal device according to an embodiment of the present disclosure. [Diagram 3] 11 is a flowchart illustrating an example of a process performed by a terminal device according to an embodiment of the present disclosure. [Figure 4] FIG. 2 is a block diagram showing a configuration example of a processing device according to an embodiment of the present disclosure. [Diagram 5] FIG. 2 is a conceptual diagram illustrating information stored in a storage device according to an embodiment of the present disclosure. [Figure 6] 11 is a flowchart illustrating a selection process of a well-being score calculation model according to an activity factor group according to an embodiment of the present disclosure. [Figure 7] FIG. 11 is a conceptual diagram illustrating an activity emotion group discrimination parameter set according to an embodiment of the present disclosure. [Figure 8] FIG. 13 is a conceptual diagram illustrating an active opposing group discrimination parameter set according to an embodiment of the present disclosure. [Figure 9] 11 is an explanatory diagram showing a derived value of a feature amount of an acceleration parameter according to an embodiment of the present disclosure. FIG. [Figure 10] FIG. 13 is a conceptual diagram illustrating a weighting table of a well-being score calculation model according to an embodiment of the present disclosure. [Figure 11] 11 is a flowchart illustrating a calculation process of a well-being score according to an embodiment of the present disclosure. [Figure 12] 11 is a flowchart illustrating a well-being score calculation process according to an embodiment of the present disclosure. [Figure 13] 1 is a conceptual diagram illustrating feedback information created by a feedback content creation unit according to an embodiment of the present disclosure. [Figure 14] FIG. 11 is a block diagram showing another form of a terminal device according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0011] Before explaining the well-being score generation system, we will first briefly explain the well-being, which is a known concept. Well-being is sometimes expressed as happiness. There are three factor axes that can measure well-being: the activity axis, the emotion axis, and the interaction (communication) axis.

[0012] The activity intensity factor, which is one of the factor elements of the activity axis, is calculated based on the acceleration intensity and the amount of activity. The stationary fixed factor, which is another factor element of the activity axis, is calculated based on the duration of stationary time.

[0013] The emotion intensity factor, which is a factor element of the emotion axis, has the content of positive or negative emotion intensity, and is calculated based on the acceleration parameters.

[0014] The opposition factor, which is a factor element of the opposition (communication) axis, indicates whether a person's opposition tendency is directed inward or outward, and is calculated based on the acceleration parameters.

[0015] In quantifying well-being as described above, data obtained from a sensor in the user's wearable device, more specifically an accelerometer, may be used, for example.

[0016] Here, it has been found that emotions toward activity (strength of positive or negative) can be broadly categorized into people who tend to feel more positive when they are active (active group MG) and people who feel more positive when they are quiet (quiet group QG). Furthermore, personality traits such as the active group MG and the quiet group QG differ in dominance in agreeableness and conscientiousness in the five-factor model used to evaluate personality in the field of psychology.

[0017] Therefore, in the well-being score generation system of the present disclosure, the well-being score is appropriately quantified by estimating the well-being score according to the activity emotion characteristic group.

[0018] FIG. 1 is a conceptual diagram showing an example of the configuration of a well-being score generating system according to an embodiment of the present disclosure.

[0019] The well-being score generation system 1 includes a terminal device 10, a processing device 20, and a storage device 30. The terminal device 10, the processing device 20, and the storage device 30 are communicatively connected to each other via a communication network 60 such as the Internet.

[0020] Typical examples of the terminal device 10 include a smartphone, a notebook computer, and a wearable device, but the types of the terminal device 10 are not limited to these.

[0021] FIG. 2 is a block diagram illustrating a configuration example of a terminal device according to an embodiment of the present disclosure.

[0022] The terminal device 10 has a general configuration including a processor, a memory, an input device, and an output device. The terminal device 10 may further include a clock 11 and a sensor 12. The terminal device 10 functionally realizes a data transmission / reception unit 13 and a feedback unit 14 by executing a program stored in the memory. The terminal device 10 may functionally realize the clock 11 and the sensor 12 by executing a program stored in the memory.

[0023] The clock 11 measures the time in the terminal device 10. The sensor 12 acquires sensor data that measures the state of the user. The data transmitter / receiver 13 transmits and receives data to and from external devices such as the processing device 20 and the storage device 30. The feedback unit 14 provides feedback to the user by presenting the user's well-being score.

[0024] The sensor data is, for example, acceleration data acquired by an acceleration sensor attached to the user's body or carried by the user.

[0025] FIG. 3 is a flowchart illustrating an example of processing performed by the terminal device 10 according to an embodiment of the present disclosure.

[0026] The clock 11 acquires the measurement time (S101). The sensor 12 acquires the user's sensor data (S102). The data transmitter / receiver 13 transmits the measurement time and the sensor data to the processing device 20 (S103). The processor of the terminal device 10 determines whether or not to end the measurement process (S104). For example, when an external event occurs, such as the user operating the terminal device 10 to press a process end button, the processor of the terminal device 10 determines to end the measurement process. If the measurement process is not to be ended (S104: NO), the process returns to step S101.

[0027] FIG. 4 is a block diagram illustrating a configuration example of a processing device according to an embodiment of the present disclosure.

[0028] The processing device 20 is a computer having a general configuration including a processor, a memory, an input device, an output device, etc. The processing device 20 functionally realizes an input unit 21, a processing unit 22, and an output unit 23 by the processor executing a program stored in the memory.

[0029] The input unit 21 has a personality characteristic input unit 211 and a sensor data input unit 212. The processing unit 22 has a personality factor score calculation unit 221, an activity factor characteristic discrimination unit 222, a well-being score calculation model selection unit 223, a sensor parameter calculation unit 224, a well-being score calculation unit 225, and a feedback content creation unit 226. The output unit 23 has a sensor parameter output unit 231, a well-being score output unit 232, and a feedback content output unit 233.

[0030] The personality characteristic input unit 211 acquires user answer information in which the user's answers to the personality test are recorded. The sensor data input unit 212 acquires sensor data of the user.

[0031] The personality factor score calculation unit 221 calculates a personality factor score, which is a score representing the characteristics of the user's personality, based on the user answer information. The activity factor characteristic discrimination unit 222 identifies the activity factor group to which the user corresponds based on the calculated personality factor score. The well-being score calculation model selection unit 223 selects a well-being score calculation model based on the activity factor group to which the user corresponds. The sensor parameter calculation unit 224 calculates the user's feature amount (sensor parameter) based on the sensor data. The well-being score calculation unit 225 calculates the well-being score by inputting the user's feature amount into the selected well-being score calculation model. The well-being score means a numerical representation of the strength of well-being. The feedback content creation unit 226 creates feedback content for the user.

[0032] The sensor parameter output section 231 outputs the feature amount (sensor parameter) of the user. The well-being score output section 232 outputs the well-being score. The feedback content output section 233 outputs the feedback content.

[0033] FIG. 5 is a conceptual diagram illustrating information stored in a storage device according to an embodiment of the present disclosure.

[0034] The storage device 30 stores a well-being score calculation model 31, a personality characteristic 32, an activity factor characteristic 33, sensor parameters 34, and a well-being score 35.

[0035] The well-being score calculation model 31 is a well-being score calculation model that calculates a well-being score using a feature quantity related to the user's physical activity as an input, and is a plurality of well-being score calculation models that are estimation models corresponding to activity factor groups that indicate a person type that appears in the mutual relationship between activity activity and positive emotion or interpersonal communicability. The activity factor groups include an active good group and a quiet good group that appear in the relationship between activity activity and positive emotion, and an active interpersonal group and a quiet interpersonal group that appear in the relationship between activity activity and interpersonal communicability. In this embodiment, the well-being score calculation model 31 has four models that correspond to the active good group, quiet good group, active interpersonal group, and quiet interpersonal group, respectively. The active good group, quiet good group, active interpersonal group, and quiet interpersonal group will be described later.

[0036] The meaning of the well-being score is as described above. The personality characteristics and the activity factor characteristics will be described later. The sensor parameters refer to the feature amounts of the user calculated based on the sensor data.

[0037] FIG. 6 is a flowchart illustrating a selection process of a well-being score calculation model according to an activity factor group according to an embodiment of the present disclosure.

[0038] The personality factor score calculation section 221 acquires the personality test result, and calculates the personality factor score based on the test result (S201). Note that a general personality test can be used here.

[0039] The activity factor characteristic discrimination unit 222 discriminates the activity factor group of the user from the personality factor score and the activity factor group discrimination model (S202).

[0040] The well-being score calculation model selection unit 223 selects a well-being score calculation model for the user based on the activity factors (S203).

[0041] The well-being score calculation model takes as input a feature related to acceleration intensity which represents the magnitude of acceleration, a feature related to activity amount which is represented by the number of zero crossing points of acceleration intensity in a specified analysis period, a feature related to activity duration which represents a period of continuous activity specified by the activity amount, and a feature related to stationary duration which represents a period of continuous inactivity specified by the activity amount.

[0042] (Acceleration strength) Acceleration intensity refers to the magnitude of body acceleration obtained by converting the raw three-axis body acceleration data, which is a three-dimensional vector measured by a sensor, using the following formula.

[0043]

number

[0044] (activity amount) The amount of activity refers to the number of times that the acceleration intensity data arranged in a time series crosses zero in a unit time (analysis period). The unit time may be, for example, one minute, but may also be a time other than one minute.

[0045] (activity and rest duration) The amount of activity is binarized based on the magnitude of the average amount of activity, and a state higher than the average is defined as an "active state," while a state lower than the average is defined as an "inactive state." The period during which this active state continues is defined as the "active duration." The period during which an "inactive state" continues is defined as the "static duration."

[0046] (Personality Test) A personality test is a known trait five-factor theory question test used in psychology to assess a person's personality. An example of a known personality test is the NEO.

[0047] (Personality factors and personality factor scores) The personality factors refer to the five factors: emotional stability (N), extraversion (E), openness (O), agreeableness (A), and conscientiousness (C). The English terms for the five factors are as follows: Emotional stability is Neuroticism. Extraversion is Extraversion. Openness is Openness. Agreeableness is Agreauleness. Conscientiousness is Conscientiousness.

[0048] The personality factor scores refer to the scores for each of the above personality factors (PS_N, PS_E, PS_O, PS_A, PS_C) calculated from the personality test results using known methods. Each score is expressed as a number from 0 to 100 points.

[0049] (Activity factor characteristics) The activity factor characteristic is a characteristic that indicates whether positive emotions or interpersonal communicability become stronger or weaker as the activity becomes more active. The activity factor characteristic includes an activity emotion group and an activity opposition group.

[0050] The active emotion group is made up of a group whose positive emotions become stronger as their activity increases (active group) and a group whose positive emotions become stronger as their activity decreases (calm group).

[0051] The activity-opposing group consists of a group whose interpersonal activity becomes stronger as their activity becomes more active (the active interpersonal group) and a group whose interpersonal activity becomes stronger as their activity becomes less active (the static interpersonal group).

[0052] (Personality traits) Personality traits are patterns or tendencies of thoughts, feelings, and actions that characterize an individual and are relatively persistent across a variety of situations. The above-mentioned active-emotion group and active-opposite group are examples of personality traits.

[0053] (Activity factor group discrimination model) The activity factor group discrimination model is a model that uses personality factor scores as input to calculate an activity emotion group score indicating the degree of correlation between activity activity and positive emotion, and an activity opposite group score indicating the degree of correlation between activity activity and interpersonal communication. The activity factor group discrimination model is stored in the storage device 30, etc. The activity emotion group score and the activity opposite group score can be defined as a linear combination formula AFS of personality factor scores as follows.

[0054] AFS=PS_N*kf_N+PS_E*kf_E+PS_O*kf_O+PS_A*kf_A+PS_C*kf_C

[0055] Let X be any single alphabetical character. kf_X indicates a set of weighting coefficients for each personality factor score. The weighting coefficients include a set of weighting coefficients ke_X for calculating the AFS for the activity emotion group, and a set of weighting coefficients kc_X for calculating the AFS for the activity opposition group.

[0056] The processing device 20 inputs the user's personality factor scores into the activity factor group discrimination model to calculate the user's activity emotion group score and activity opposition group score.

[0057] 7 is a conceptual diagram illustrating an active emotion group discrimination parameter set according to an embodiment of the present disclosure. The active emotion group discrimination parameter set includes the set of weighting coefficients ke_X for the active emotion group described above and the active emotion group threshold AESth as values.

[0058] 8 is a conceptual diagram illustrating an example of an activity counter group discrimination parameter set according to an embodiment of the present disclosure. The activity counter group discrimination parameter set includes a set of weight coefficients kc_X for the above-mentioned activity counter group and an activity counter group threshold value ACSth as values.

[0059] In step S202 of FIG. 6, the activity factor characteristic discrimination unit 222 discriminates the activity factor group of the user from the personality factor scores (PS_N, PS_E, PS_O, PS_A, PS_C) and the activity factor group discrimination model.

[0060] For example, the activity factor characteristic discrimination unit 222 discriminates the active emotion group as a pro-active group if the AFS (active emotion group score) for the active emotion group calculated using the above formula is greater than a threshold AESth, and discriminates the active emotion group as a sedentary group if it is smaller. That is, the activity factor characteristic discrimination unit 222 of the processing device 20 judges that the user belongs to the pro-active group, which is a type of person whose positive emotions become stronger as the activity becomes more active, if the active emotion group score is smaller than the threshold, and judges that the user belongs to the sedentary group, which is a type of person whose positive emotions become stronger as the activity becomes less active, if the active emotion group score is smaller than the threshold.

[0061] In addition, the activity factor characteristic discrimination unit 222 discriminates the active counter group as the active interpersonal group if the AFS (active counter group score) for the active counter group calculated using the above formula is greater than a threshold ACSth, and discriminates the active counter group as the static interpersonal group if it is smaller. That is, the activity factor characteristic discrimination unit 222 of the processing device 20 judges that the user corresponds to the active interpersonal group, which is a person type whose interpersonal communicativeness becomes stronger as the activity becomes more active, if the active emotion group score is smaller than the threshold, the user corresponds to the static interpersonal group, which is a person type whose interpersonal communicativeness becomes stronger as the activity becomes less active.

[0062] (Derived value of acceleration parameter feature) FIG. 9 is an explanatory diagram showing a derived value of a feature of an acceleration parameter according to an embodiment of the present disclosure. The feature of an acceleration parameter input to the well-being score calculation model includes acceleration intensity, activity amount, activity duration, and rest duration. The derived values ​​of these feature amounts can also be the feature amounts input to the well-being score calculation model. More specifically, the well-being score calculation model inputs the skewness, kurtosis, and median of each of the acceleration intensity, activity amount, activity duration, and rest duration, and the probability density of each period obtained by dividing the analysis period into a plurality of equal parts. Note that d1-dn in the figure will be explained. dn means the probability density for the class width of the section width that is divided into n equal parts for the entire section of the acceleration parameter intensity from which noise has been removed.

[0063] (Well-being score calculation model) The well-being score calculation model for calculating the well-being score may be implemented as a linear model, a non-linear model, or a machine learning model. In this embodiment, an example of a well-being score calculation model based on a linear model is shown.

[0064] 10 is a conceptual diagram showing a weighting table of a well-being score calculation model according to an embodiment of the present disclosure. In the case of a linear model, the well-being score is calculated based on the sum of values ​​obtained by multiplying the feature amount by a coefficient, that is, based on a linear combination.

[0065] The types of feature amount include four types: acceleration intensity, activity amount, activity duration, and stillness duration. For each of these types, the above-mentioned skewness, kurtosis, median, and d1-dn values ​​can be calculated, and the calculated values ​​can be treated as feature amounts of acceleration parameters. The well-being score calculation unit 225 calculates the well-being score WBS by adding the results of multiplying the feature amount by a weighting coefficient using a well-being score calculation model that is a linear model. That is, the processing device 20 calculates the well-being score as a value obtained by linearly combining values ​​based on the input of the well-being score calculation model.

[0066] For example, if the type of acceleration parameter is expressed by the subscript p and the type of feature is expressed by the subscript f, then the value V of feature f of acceleration parameter p is pf and its weighting factor kg pf Using this, the well-being score WBS can be expressed by the following formula.

number

[0067] Weighting factor kg pf The value of is different for each activity factor group. That is, the weight coefficient set kmp is unique for each of the activity group, the rest group, the activity group, and the rest group. pf Set, kqp pf Set, kmo pf Set, and kmi pf Given a set, the values ​​of each coefficient are stored in the weighting table shown in FIG.

[0068] For example, when calculating the well-being score of the emotion intensity factor (well-being emotion intensity score) for a user in the good behavior group, the calculation formula for the score is as follows:

[0069]

number

[0070] Similarly, when the well-being score of the allotropic factor (well-being allotropic score) is calculated for a user of the moving-to-person group, the formula for calculating the score is as follows:

number

[0071] FIG. 11 is a flowchart illustrating a process for calculating a well-being score according to an embodiment of the present disclosure.

[0072] The well-being score calculation section 225 reads the well-being score calculation model selected by the well-being score calculation model selection section 223 (S301). The well-being score calculation section 225 acquires the measurement time and sensor data from the terminal device 10 (S302).

[0073] The well-being score calculation section 225 performs a well-being score calculation process (S303). Details of the well-being score calculation process will be described later with reference to Fig. 12. The well-being score calculation section 225 records the calculated well-being score in the storage device 30 (S304).

[0074] The well-being score calculation unit 225 determines whether or not to end the calculation process of the well-being score (S305). For example, when the user operates the terminal device 10 to press a process end button and an external event indicating the end of the process occurs, the well-being score calculation unit 225 determines to end the calculation process of the well-being score. If the process is not to be ended (S305: NO), the process returns to step S302.

[0075] FIG. 12 is a flowchart illustrating a well-being score calculation process according to an embodiment of the present disclosure.

[0076] The well-being score calculation unit 225 calculates acceleration parameters (S401). For example, the well-being score calculation unit 225 calculates the skewness, kurtosis, median, and d1-dn, which are feature quantities of the acceleration parameters, based on the sensor data acquired in step S302 of FIG.

[0077] The well-being score calculation section 225 calculates the well-being score using the above-mentioned acceleration parameters and the well-being score calculation model (S402).

[0078] FIG. 13 is a conceptual diagram illustrating feedback information created by the feedback content creation unit 226 according to an embodiment of the present disclosure.

[0079] The feedback information includes a measurement date, a measurement time, and a well-being score. The measurement date and the measurement time are values ​​based on the measurement time transmitted by the terminal device 10 to the processing device 20 in step S103 of Fig. 3. The well-being score includes a well-being emotion intensity score and a well-being opposite score.

[0080] The feedback content output unit 233 of the processing device 20 outputs the feedback information created by the feedback content creation unit 226. The output feedback information is received by the data transmission / reception unit 13 of the terminal device 10, and the feedback unit 14 presents the feedback information to the user.

[0081] Fig. 14 is a block diagram showing another form of a terminal device according to an embodiment of the present disclosure. Among the components of the terminal device 10 shown in Fig. 2, the sensor 12 is connected to the terminal device 10 as a sensor terminal 50 which is an external device seen from the terminal device 10.

[0082] (Improved accuracy based on user input information) In the above, the well-being score is calculated based on the acceleration data based on the sensor data acquired by the sensor 12. The accuracy of the well-being score may be further improved by further acquiring the user's situation, feelings, and the like.

[0083] For example, an input function is added to the terminal device 10. The terminal device 10 further includes a data input unit. The user inputs the well-being score for each arbitrary period, such as one week. The data transmission / reception unit 13 of the terminal device 10 transmits information on the input well-being score to the processing device 20.

[0084] The processing device 20 corrects the well-being score calculation model so that the difference between the experience score acquired from the terminal device 10 and the well-being score calculated using the well-being score calculation model in use is minimized. For example, in the case of a well-being score calculation model that is a linear model, the currently used weighting coefficient set is set as an initial value. Then, the weighting coefficient set is updated using a minimum optimization algorithm so that the difference from the experience score is minimized. In the case that the well-being score calculation model is a machine learning model, additional learning is performed based on error backpropagation using the experience score as training data, and the weighting coefficient set is updated so that the difference from the experience score is minimized.

[0085] This makes it possible to calculate a highly accurate well-being score that is tailored to the user.

[0086] (Utilizing vital sensor information) In the above, the well-being score is calculated based on the acceleration data based on the sensor data acquired by the sensor 12. In addition to the acceleration data, vital sensor information (heart rate, surface skin conductance) may be utilized.

[0087] The terminal device 10 further includes a vital sensor, or the terminal device 10 is communicably connected to a vital sensor. The data transmitter / receiver 13 transmits vital sensor information, such as the user's heart rate and surface skin conductivity, acquired by the vital sensor to the processing device 20.

[0088] The processing device 20 is provided with a heart rate factor group discrimination section and a sweat factor group discrimination section similar to the activity factor characteristic discrimination section 222. The heart rate factor group discrimination section performs discrimination using a heart rate factor group discrimination model. The sweat factor group discrimination section performs discrimination using a sweat factor group discrimination model. The heart rate factor group discrimination model and the sweat factor group discrimination model are each defined by a linear combination formula AFS of personality factor scores. However, a set of weighting coefficients specific to each model is used. As with the activity factor characteristic discrimination section 222, the heart rate factor group and the sweat factor group are each discriminated depending on whether the calculated AFS value is greater than a threshold value.

[0089] The processor of the processing device 20 calculates a well-being score for each heart rate from the heart rate measured by the vital sensor using a well-being score calculation model specific to the heart rate factor group, similar to the acceleration. Similarly, the processor of the processing device 20 calculates a well-being score for each sweat factor from the surface skin conductivity measured by the vital sensor using a well-being score calculation model specific to the sweat factor group.

[0090] As a result of the processor of the processing device 20 performing the above processing, a well-being score based on acceleration, a well-being score based on the heart rate factor group, and a well-being score based on the sweat factor group coexist. The processor of the processing device 20 performs a weighted average on the calculated multiple well-being scores, and calculates the weighted average as the final well-being score.

[0091] This makes it possible to calculate a highly accurate well-being score by taking into account not only acceleration but also heart rate and sweat levels.

[0092] (Sleep state based correction) The state of sleep is considered to have a significant effect on well-being. Therefore, the processor of the processing device 20 may correct the well-being score to be calculated based on the state of sleep.

[0093] The terminal device 10 includes a data input unit. A user inputs information indicating a sleep state into the data input unit of the terminal device 10. The information indicating the sleep state may be, for example, the sleeping hours of the previous day. The data transmission / reception unit 13 of the terminal device 10 transmits the information on the sleeping hours acquired from the user to the processing device 20.

[0094] The processor of the processing device 20 calculates the ratio between the received sleep time and the average sleep time, and uses the ratio between the sleep time and the average sleep time as the sleep state information SS indicating the sleep state.

[0095] The processor of the processing device 20 calculates the sleep correction factor based on the following formula: Sleep correction factor = SS × k sleep In addition, k sleep is a sleep correction constant for the sleep correction factor.

[0096] The processor of the processing device 20 multiplies the well-being score calculated in step S402 of FIG. 12 by the sleep correction coefficient to calculate a well-being score that takes into account the influence of the user's sleep state.

[0097] The calculation process of the sleep state information SS may be performed by the processor of the terminal device 10. The processor of the processing device 20 may multiply the well-being score based on the heart rate factors or the well-being score based on the sweating factors by a sleep correction coefficient, or may multiply a weighted average value of a plurality of well-being scores by the sleep correction coefficient.

[0098] This makes it possible to calculate a highly accurate well-being score that takes into account sleep state.

[0099] (environmental factors) The sensor data may include data on environmental factors that measure the state of the environment around the user, and the well-being score calculation model may calculate the well-being score using the feature amount and the environmental factors as input.

[0100] The environmental information sensor measures the state of the environment around the user, such as temperature, humidity, air pressure, and illuminance. The environmental information sensor is provided in the terminal device 10 or the processing device 20. Alternatively, the environmental information sensor is connected to the terminal device 10 or the processing device 20 so as to be able to communicate with them.

[0101] The processor of the processing device 20 acquires environmental information directly or indirectly from the environmental information sensor. The processor of the processing device 20 calculates the following values ​​based on the acquired environmental information. Press now : Atmospheric pressure at the time of measurement Hum now : Humidity at the time of measurement Temp now : Temperature at the time of measurement Press 24hAgo : Atmospheric pressure 24 hours prior to measurement Hum 24hAgo : Humidity 24 hours before measurement, Temp 24hAgo : Temperature 24 hours prior to measurement Illum mean : The number of hours of sunlight on that day

[0102] The processor of the processing device 20 calculates the correction coefficient ES based on the following formula: weather , E.S. heatstroke , and E.S. SAD Calculate.

[0103] ES weather =(Press now / Press 24hAgo )×k press ES heatstroke =(Hum now / Hum 24hAgo )×k hum +(Temp now / Temp 24hAgo )×k temp ES SAD =Illum mean ×k illum

[0104] k press is a correction constant for atmospheric pressure. k hum is a correction constant for humidity. k temp is a correction constant for temperature. k illum is a correction constant for the average sunshine duration correction factor.

[0105] ES weather is a correction coefficient indicating the applicability of meteorological illness. ES heatstrokeis a correction coefficient indicating the susceptibility to heat stroke. ES SAD is a correction coefficient indicating the relevance of seasonal affective disorder.

[0106] The processor of the processing device 20 applies a correction coefficient ES to the well-being score calculated in step S402 of FIG. weather , E.S. heatstroke , and E.S. SAD The well-being score taking into account the influence of the user environment is calculated by multiplying the above by 1.

[0107] The processor of the processing device 20 calculates the correction coefficient ES weather , E.S. heatstroke , and E.S. SAD may be multiplied by the well-being score based on the heart rate factors or the well-being score based on the sweat factors, or may be multiplied by a weighted average of multiple well-being scores.

[0108] This makes it possible to calculate a highly accurate well-being score that takes into account the user's environment.

[0109] The above-described embodiments of the present invention are illustrative examples of the present invention, and are not intended to limit the scope of the present invention to these embodiments. A person skilled in the art can implement the present invention in various other forms without departing from the scope of the present invention. In addition, the technical scope of the above-described embodiments includes the matters shown in the following appendices. However, the present invention is not limited to the following appendices. However, the matters included in the above-described embodiments are not limited to those appended.

[0110] As described above, the well-being score generation system generates a well-being score that numerically represents the strength of a user's well-being from sensor data that measures the user's state, A storage device and a processing device are included. The storage device is a well-being score calculation model that calculates a well-being score using a feature quantity related to a user's physical activity as an input, the well-being score calculation model being an estimation model corresponding to a group of activity factors that indicate a person type that appears in the mutual relationship between activity vibrancy and positive emotions or interpersonal communicability, The processing device includes: Acquire user answer information in which the user's answers to the personality test are recorded; A personality factor score, which is a score representing the characteristics of the user's personality, is obtained based on the user's answer information; Identifying a group of activity factors that the user is eligible for based on the personality factor scores; Acquire sensor data of the user; Calculate user features based on sensor data, The user selects a well-being score calculation model based on the corresponding activity factor group, The user's feature amount is input into the selected well-being score calculation model to calculate the well-being score. As a result, a well-being score calculation model is prepared in advance for each activity factor group, and the well-being score is calculated using the well-being score calculation model selected based on the activity factor group, making it possible to appropriately quantify the well-being of a variety of individuals.

[0111] The sensor data is acceleration data acquired by an acceleration sensor attached to the user's body or carried by the user. As a result, well-being is calculated based on acceleration data, making it possible to measure an individual's well-being in ordinary daily life using an acceleration sensor that is easy for users to wear or carry.

[0112] The well-being score calculation model includes four models corresponding to the active group, which is a type of person whose positive emotions grow stronger as their activity increases, the quiet group, which is a type of person whose positive emotions grow stronger as their activity decreases, the active interpersonal group, which is a type of person whose interpersonal communicativeness grows stronger as their activity increases, and the quiet interpersonal group, which is a type of person whose interpersonal communicativeness grows stronger as their activity decreases. The storage device is Further storing an activity factor group discrimination model that uses the personality factor scores as input to calculate an activity emotion group score indicating the degree of correlation between activity activity and positive emotion and an activity opposition group score indicating the degree of correlation between activity activity and interpersonal communication ability; The processing device includes: Calculating the user's activity emotion group score and activity opposition group score using the user's personality factor score as an input to an activity factor group discrimination model; If the active emotion group score is greater than a predetermined threshold, the user is determined to belong to an active group. If the active emotion group score is less than the threshold, the user is determined to belong to a quiet group. If the activity interaction group score is greater than a predetermined threshold, the user is determined to belong to the active interaction group, and if the activity emotion group score is less than the threshold, the user is determined to belong to the static interaction group. This makes it possible to calculate a more appropriate well-being score according to the user's activity emotion group and activity opposite group.

[0113] The well-being score calculation model takes as input a feature related to acceleration intensity which represents the magnitude of acceleration, a feature related to activity amount which is represented by the number of zero crossing points of acceleration intensity in a specified analysis period, a feature related to activity duration which represents a period of continuous activity specified by the activity amount, and a feature related to stationary duration which represents a period of continuous inactivity specified by the activity amount. This makes it possible to calculate a more appropriate well-being score by taking into account the amount of activity, the duration of activity, and the duration of stillness in addition to the acceleration intensity.

[0114] The well-being score calculation model inputs the skewness, kurtosis, and median of acceleration intensity, activity amount, activity duration, and stationary duration, as well as the probability density of each period obtained by dividing the analysis period into multiple equal parts. This makes it possible to calculate a more appropriate well-being score that takes into account the derived values ​​of the features, such as skewness, kurtosis, and median, as well as the probability density of each period into which the analysis period is divided equally.

[0115] The processing device calculates the well-being score as a linear combination of values ​​based on inputs of a well-being score calculation model. This makes it possible to linearly combine the various feature amounts described above to calculate a well-being score that takes various factors into consideration.

[0116] The sensor data includes data on environmental factors that measure the state of the environment around the user; The well-being score calculation model calculates a well-being score using the feature amounts and environmental factors as input. This makes it possible to calculate a highly accurate well-being score that takes into account the user's environment.

[0117] A well-being score generating method for generating a well-being score that numerically represents an intensity of a user's well-being from sensor data that measures a state of the user by a system having at least a storage device and a processing device, comprising: The computer a well-being score calculation model that calculates a well-being score using a feature quantity related to a user's physical activity as an input, the well-being score calculation model being an estimation model corresponding to a group of activity factors that indicate a person type that appears in the mutual relationship between activity vibrancy and positive emotions or interpersonal communicability, Acquire user answer information in which the user's answers to the personality test are recorded; A personality factor score, which is a score representing the characteristics of the user's personality, is obtained based on the user's answer information; Identifying a group of activity factors that the user is eligible for based on the personality factor scores; Acquire sensor data of the user; Calculate user features based on sensor data, The user selects a well-being score calculation model based on the corresponding activity factor group, The well-being score is calculated by inputting the user's feature amount into the selected well-being score calculation model. As a result, a well-being score calculation model is prepared in advance for each activity factor group, and the well-being score is calculated using the well-being score calculation model selected based on the activity factor group, making it possible to appropriately quantify the well-being of a variety of individuals. [Explanation of symbols]

[0118] 1...well-being score generation system, 10...terminal device, 11...clock, 12...sensor, 13...data transmission / reception unit, 14...feedback unit, 20...processing device, 21...input unit, 211...personality characteristic input unit, 212...sensor data input unit, 22...processing unit, 221...personality factor score calculation unit, 222...activity factor characteristic discrimination unit, 223...well-being score calculation model selection unit, 224...sensor parameter calculation unit, 225...well-being score calculation unit, 226...feedback content creation unit, 23...output unit, 231...sensor parameter output unit, 232...well-being score output unit, 233...feedback content output unit, 30...storage device, 31...well-being score calculation model, 32...personality characteristic, 33...activity factor characteristic, 34...sensor parameter, 35...well-being score, 50...sensor terminal, 60...communication network

Claims

1. A well-being score generation system that generates a well-being score that numerically represents an intensity of a user's well-being from sensor data that measures a user's state, comprising: A storage device and a processing device are included. The storage device includes: a well-being score calculation model that calculates a well-being score using a feature quantity related to a user's physical activity as an input, the well-being score calculation model being an estimation model corresponding to a group of activity factors that indicate a person type that appears in the mutual relationship between activity vibrancy and positive emotions or interpersonal communicability, The processing device includes: acquiring user answer information in which answers of the user to a personality test are recorded; acquiring a personality factor score, which is a score representing a personality characteristic of the user, based on the user answer information; Identifying a group of activity factors to which the user corresponds based on the personality factor scores; acquiring sensor data of the user; Calculating a feature amount of the user based on the sensor data; Selecting the well-being score calculation model based on a group of activity factors corresponding to the user; inputting the feature amount of the user into the selected well-being score calculation model to calculate a well-being score; Well-being score generation system.

2. The sensor data is acceleration data acquired by an acceleration sensor attached to the user's body or carried by the user. The well-being score generation system according to claim 1 .

3. The well-being score calculation model includes four models corresponding to a person type of an active group whose positive emotions become stronger as their activity becomes more active, a quiet group whose positive emotions become stronger as their activity becomes less active, an active interpersonal group whose interpersonal communicativeness becomes stronger as their activity becomes more active, and a quiet interpersonal group whose interpersonal communicativeness becomes stronger as their activity becomes less active, The storage device includes: Further storing an activity factor group discrimination model that uses the personality factor scores as input to calculate an activity emotion group score indicating the degree of correlation between activity activity and positive emotion and an activity opposing group score indicating the degree of correlation between activity activity and interpersonal communication ability, The processing device includes: Calculating an activity emotion group score and an activity opposition group score for the user using the personality factor score of the user as an input to the activity factor group discrimination model; If the active emotion group score is greater than a predetermined threshold, the user is determined to belong to the active group. If the active emotion group score is less than the threshold, the user is determined to belong to the quiet group. If the activity interaction group score is greater than a predetermined threshold, the user is determined to belong to the active interaction group, and if the activity emotion group score is less than the threshold, the user is determined to belong to the static interaction group. The well-being score generation system according to claim 1 .

4. The well-being score calculation model receives as input a feature amount related to acceleration intensity representing the magnitude of acceleration, a feature amount related to an activity amount represented by the number of zero crossing points of the acceleration intensity in a predetermined analysis period, a feature amount related to an activity duration representing a period during which an active state specified by the activity amount continues, and a feature amount related to a stillness duration representing a period during which an inactive state specified by the activity amount continues. The well-being score generation system according to claim 1 .

5. The well-being score calculation model inputs the skewness, kurtosis, and median of the acceleration intensity, the activity amount, the activity duration, and the stationary duration, and the probability density of each period obtained by equally dividing the analysis period into a plurality of periods. The well-being score generating system according to claim 4 .

6. The processing device calculates the well-being score as a linear combination of values ​​based on an input of the well-being score calculation model. The well-being score generating system according to claim 4 .

7. The sensor data includes data on environmental factors that measure a state of an environment around the user; the well-being score calculation model calculates a well-being score using the feature amount and the environmental factor as inputs; The well-being score generation system according to claim 1 .

8. A well-being score generation method for generating a well-being score that numerically represents an intensity of a user's well-being from sensor data that measures a user's state, comprising: The computer a well-being score calculation model that calculates a well-being score using a feature quantity related to a user's physical activity as an input, the well-being score calculation model being an estimation model corresponding to a group of activity factors that indicate a person type that appears in the mutual relationship between activity vibrancy and positive emotions or interpersonal communicability, acquiring user answer information in which answers of the user to a personality test are recorded; acquiring a personality factor score, which is a score representing a personality characteristic of the user, based on the user answer information; Identifying a group of activity factors to which the user corresponds based on the personality factor scores; acquiring sensor data of the user; Calculating a feature amount of the user based on the sensor data; Selecting the well-being score calculation model based on a group of activity factors corresponding to the user; inputting the feature amount of the user into the selected well-being score calculation model to calculate a well-being score; How well-being scores are generated.

Images