JP2020010893A - Activity classification method and activity classification device - Google Patents

Abstract

To provide an activity classification method that allows information on a behavior of an object person to be acquired from an activity pattern with high accuracy even for an object living with a plurality of activity patterns.SOLUTION: Activities are classified by an activity classification method including an information acquisition step (Step S102) for acquiring self activity information indicating an amount of activity of an object person S for each time zone a plurality of times for the object person, and a group generation step (Step S104 and Step S105) for classifying a plurality of pieces of the self activity information or groups of the self activity information according to similarities to each other, and generating a plurality of groups.SELECTED DRAWING: Figure 6

Description

  The present invention relates to an activity classification method and an activity classification device for classifying human activity patterns.

  It is known that there is an association between the state of activity of an organism, such as a human, and the physical and mental state of this organism. For this reason, information indicating the state of biological activity is collected, and a change in the state of the subject is detected at an early stage based on the collected information. An invention for collecting data on the activity of a subject and detecting an abnormality in living behavior is described in Patent Document 1, for example. The living activity monitoring system described in Patent Literature 1 determines whether there is an abnormality of a single person from a disordered activity pattern. The invention described in Patent Literature 1 is suitable for detecting a disturbance in an activity pattern when a subject performs an activity in the same pattern every day.

JP 2000-57462 A

However, the activity pattern is different between a certain day and a non-work day when the subject is working while being subjected to social restrictions such as work and learning. For this reason, when the invention described in Patent Document 1 is applied to a generation having a job (hereinafter referred to as “occupational generation”), the variation in the activity becomes large, and the reliability of information obtained from the activity pattern increases. Is reduced.
The present invention has been made in view of such a point, and even for a target living under social constraints, an activity that can obtain information on the behavior of the target person from the activity pattern with high accuracy. The present invention relates to a classification method and an activity classification device.

  The activity classification method of the present invention is an information acquisition step of acquiring self-activity information indicating the amount of activity of a subject for each time period a plurality of times for the subject, and a plurality of the self-activity information or a group of the self-activity information. And a group creation step of creating a plurality of groups by classifying the groups according to their similarities.

  Further, the activity classification device of the present invention is an information acquisition unit that acquires self activity information indicating the amount of activity of the subject for each time zone a plurality of times for the subject, a plurality of the self activity information, or the self activity information And a group creating unit that creates a plurality of groups by classifying the groups according to their similarities.

  The present invention can provide an activity classification method and an activity classification device that can obtain information on the behavior of a target person from an activity pattern with high accuracy even for a target living with a plurality of activity patterns.

It is a schematic diagram for explaining the concept of the activity classification method and the activity classification device of the present embodiment. FIG. 2 is a block diagram for explaining functions of the server device shown in FIG. 1. (A) to (e) are diagrams showing self-activity information and self-activity patterns extracted from the self-activity information. (A) to (d) are diagrams showing self-activity information for comparison with FIG. 3 and self-activity patterns extracted from this self-activity information. FIG. 3 is a diagram showing a state in which a message created by the mind and body information creation unit shown in FIG. 2 is displayed on a smartphone. FIG. 3 is a diagram illustrating a flowchart for describing an activity classification method executed by the server device illustrated in FIGS. 1 and 2. (A) to (f) are diagrams showing comparative activity information of a plurality of other subjects grouped.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In all the drawings, the same components are denoted by the same reference numerals, and redundant description will be omitted as appropriate.

<Overview>
Prior to the description of the embodiments of the present invention, the concept of the present invention will be described.
FIG. 1 is a schematic diagram for explaining the concept of an activity classification method and an activity classification device according to a first embodiment and a second embodiment of the present invention (hereinafter, both are also referred to as “this embodiment”). . The activity classification method according to the present embodiment is realized by the server device 1 which is an activity classification device. The server device 1 acquires the activity amount of the target person S and the time period during which the activity was performed. Then, the acquired activity amount and the time period during which the activity was performed are totaled for each subject S. In the present embodiment, information indicating the relationship between the activity amount of the target person S and the time period during which this activity was performed is referred to as “self activity information”.
The self-activity information may be, for example, the one that acquires the number of steps of walking using a known pedometer and associates the number of steps with the time to obtain a time zone of walking. Further, as the pedometer, for example, a pedometer incorporated in a smartphone or the like may be used. Further, the amount of the activity is not limited to the number of steps, and may be a value obtained by acquiring the moving distance of the subject S using a GPS (Global Positioning System) or the like. In the case of acquiring the amount of exercise based on the moving distance, naturally, a movement at a speed higher than the walking speed of a human is excluded from walking as movement by a vehicle.
Note that the present embodiment is not limited to one that counts the number of steps using a pedometer. For example, an acceleration sensor may be attached to a subject, and the activity amount of the subject may be detected based on the acceleration generated in the subject.

Further, in the present embodiment, information on the intensity of the activity together with the amount of the activity may be included in the self-activity information. For example, when the pedometer uses a three-axis acceleration sensor, the activity intensity may be detected by the intensity of acceleration applied to at least one axis in three directions. In the case where the activity amount is measured by the movement distance, the activity intensity may be detected by the movement speed.
In the present embodiment, the self-activity information is not limited to a combination of the activity amount or intensity and the activity time zone. The self-activity information may be any combination of information on the activity and information on the time or time at which the activity was performed. In addition, the amount of the activity is not limited to a numerical value obtained from the measurement of the number of steps, but is represented by a processed numerical value such as a ratio of the number of steps in the number of steps taken by the subject S in one day. There may be.

  The server device 1 shown in FIG. 1 constructs a system together with the smartphone P possessed by the target person S. The smartphone P has a pedometer function as an existing function. Then, the number of steps counted by the pedometer function is transmitted to the server device 1 using the communication function of the smartphone P. The name, age, sex, identification information of the smartphone P, and the like of the target person S are registered in the server device 1 in advance, and the number of steps is input to the server device 1 as data of the target person S together with the reception time of the information on the number of steps. . The server device 1 creates self-activity information from the input information. The self activity information is stored in a database 7 (denoted as DB in the figure).

In the example shown in FIG. 1, the target person S is a company employee who commutes to the company by train. It is assumed that the target person S goes to work from Monday to Friday and works from 9:00 to 18:00. In such an example, the subject S starts walking toward the nearest station at, for example, 8:00 in scene A in FIG. Then, he stops walking on the train on the way and starts walking again after arriving at the nearest station of the company. In addition, the target person S starts work in the scene B. In this example, the subject S takes a work mode centered on desk work, and the number of steps of the subject S walking in the scene B is smaller than that in the scene A. Further, in the scene C, the subject S walks from the company to the nearest station of the company, gets on the train, moves to the nearest station at home, and then walks from the nearest station to home.
In the above scene, the smartphone P transmits information indicating the number of steps of the target person S to the server device 1 every few seconds, for example, and the server device 1 adds and stores the transmitted number of steps. The number of steps is added, for example, every hour. In such a case, the added value of the 24 steps is stored in the server device 1 during one day.

  As described above, when the target person S is a job generation who works for the company, the approximate number of steps on the working day (weekday) of the target person S is determined by social constraints such as attendance and working hours. . Then, the number of steps of the holiday of the subject S is different from that of weekdays due to different social restrictions. In addition, when the target person S regularly attends a lesson, a sports gym, or the like after leaving work or before going to work, the number of steps of the target person S differs between a day on which there is no learning and a day on which there is no lesson. It is conceivable that the number of steps on the day of going to a lesson or a sports gym will be approximately a constant value due to social restrictions such as the start and end times of the lesson and the use time zone of the gym.

As described above, when the subject S lives under social constraints, when the activity (action) information is recorded over a long period of time and the change is detected as described in Patent Document 1, The variability of the recorded self-activity information increases, and the reliability of the detection result decreases. Therefore, in the present embodiment, the self-activity information of the subject S is accumulated, and the accumulated self-activity information is classified into a plurality of groups based on the degree of similarity with each other.
According to the present embodiment, for example, a change in the activity of the subject S can be detected with high accuracy every weekday or every holiday. Further, for this reason, it is possible to appropriately determine the relevance between the change in the activity and the physical or mental state of the subject S. Further, the target person S can be compared with the activity information of the other person targeted by the target person S, and advice can be given for the target person S to approach the target. Furthermore, by detecting a change with time of the activity of the subject S, the physical strength, energy, and lifestyle of the subject S can be reviewed.

[First embodiment]
FIG. 2 is a block diagram for explaining functions of the server device 1 shown in FIG. In the first embodiment, the server device 1 corresponds to an activity classifying device. The server device 1 collects self-activity information from the receiving device 3 and stores it in the database 7. However, the server device 1 may classify the self-activity information collected in advance. The database 7 may be connected to the server device 1 via a network. The server device 1 includes a calculation unit 10, a receiving device 3, and a transmitting device 5, and the receiving device 3 receives an activity amount from the smartphone P of the subject S, for example, every two seconds. The input unit 11 of the server device 1 sums the received activity amounts for a predetermined time width, and receives a reception time of the first activity amount and a reception time of the last activity amount of the plurality of combined activity amounts. And the time interval (time zone). In the first embodiment, information composed of the summed amount of activity and the time zone associated with this amount of activity is referred to as self-activity information. The server device 1 of the first embodiment groups the self-activity information into a plurality of groups, creates mental and physical information of the subject S, and transmits the information to the smartphone P of the subject S from the transmission device 5. Note that the mental and physical information is information relating to at least one of the body and the soul, and not only points out the physical or mental state of the subject S, but also information that advises such a state in a preferable direction. Can also be included.

  In order to fulfill the above functions, the server device 1 has the receiving device 3 which is an information acquisition unit that acquires the self activity information indicating the amount of the activity of the target person S for each time zone a plurality of times for one target person S. ing. In addition, the calculation unit 10 of the server device 1 is a group creation unit that is a group creation unit that creates a plurality of groups by classifying a plurality of self-activity information based on their similarity or a group of the self-activity information based on their similarity. It has a part 13. Such a server device 1 executes an information acquisition step of acquiring self-activity information a plurality of times, and a group creation step of creating a plurality of groups by classifying a plurality of self-activity information based on their similarities.

In the above description, the first embodiment can be applied to a subject to determine the physical condition of the organism and the stress applied to the subject. Further, in the above description, in the first embodiment, the activity amount is acquired at regular time intervals during a day, and this is performed over a plurality of days. For this reason, “acquiring a plurality of times” in the first embodiment refers to acquiring the activity information a plurality of times for one subject S. The acquisition of the activity information a plurality of times may be performed, for example, every hour. In the first embodiment, the activity information is obtained from a plurality of subjects.
In addition, the phrase “self activity information” is to distinguish the activity information of the subject S for whom the activity information is to be obtained from the activity information of another person compared with the subject S. Here, the other person may be any person other than the target person S as long as it is a target person for acquiring activity information. As will be described later, the activity information of the other person compared with the self activity information is also referred to as “comparison activity information”.

  The similarity of the self-activity information indicates the degree of similarity in the amount of activity and the time zone in which the activity was performed among a plurality of pieces of self-activity information. For example, when the activity amount of the subject S is acquired on a plurality of days (day I, day II, day III, and day IV) and the self activity information is created by associating the activity amount with the time zone, Assume that S mainly performs an activity (walk) in the morning and hardly walks in the afternoon. In addition, it is assumed that the subject S mainly walks hardly in the morning on the day II and walks from the afternoon to the night. Further, it is assumed that the subject S mainly walks in the morning on day III and hardly walks in the afternoon. In such a case, in the first embodiment, the grouping is performed such that the activity information on day III has the same group as the activity information on day I with priority over the activity information on day II.

  Hereinafter, each configuration of the server device 1 illustrated in FIG. 2 will be described. The server device 1 includes an arithmetic unit 10, a receiving device 3, a transmitting device 5, and a database 7 connected to the arithmetic unit 10. The arithmetic unit 10, the receiving device 3, and the transmitting device 5 can be configured using a general-purpose computer. The receiving device 3 and the transmitting device 5 may be provided separately or may be integrated. The database 7 may be connected to the operation unit 10 via a cable, or may be stored in a recording medium connectable to the operation unit 10. Further, the database 7 may be connected to the arithmetic unit 10 via a network.

(Receiving device and transmitting device)
The receiving device 3 and the transmitting device 5 may be devices that can wirelessly transmit and receive signals to and from a communication device such as the smartphone P owned by the subject S. In the first embodiment, the activity amount of the subject S is transmitted as the number of steps from the smartphone P having the pedometer function to the receiving device 3, but the first embodiment is not limited to such an example. In the first embodiment, for example, the subject S may record the number of steps counted by a pedometer, input the same to a device having a communication function, and transmit the communication device to the receiving device 3.

(Calculation unit)
The calculation unit 10 includes an input unit 11, a constraint information addition unit 12, a group creation unit 13, an activity pattern extraction unit 14, a mind and body information creation unit 15, and an output unit 16. The arithmetic unit 10 illustrated in FIG. 2 is a combination of known hardware of a computer and a program that is software for operating the hardware. The hardware includes, for example, a CPU (Central Processing Unit) that controls and controls the arithmetic unit 10, a working memory used for processing by the CPU, and a memory used for storing data. Software is a program that operates on a CPU and controls hardware to realize each function.

(Input section)
The input unit 11 of the arithmetic unit 10 accumulates the received amount of self-activity for a predetermined time, and associates the accumulated amount of self-activity with the time period from the reception start time to the reception end time of the activity amount. This is a program for activity information.

(Constraint information addition section)
The constraint information adding unit 12 has a function of associating the self-activity information with the constraint information on the constraint on the behavior of the subject S when acquiring the self-activity information. Here, the constraint information refers to, for example, information related to social activities such as work, learning, volunteer activities, and the like (hereinafter, referred to as “work, etc.”) of the subject S. The social behavior refers to a behavior that the subject S cannot arbitrarily determine and change, or an arbitrarily difficult to determine and change. The constraint information includes information indicating that the social constraint is strong and information indicating that the social constraint is weak. The social constraint refers to a social rule that has a compelling force on the target person S, and a strong constraint refers to a strong coercive force. In addition, a weak constraint indicates that the forcing force is weak. For example, "work" is enforced by a force such as "arriving at the work place by the starting time on weekdays" or "being at the work place during working hours". In addition, for "going to a gym", for example, a compulsory force such as "arriving at an office during business hours of the gym" or "arriving at the facility by the reserved time of the facility" is applied. In the first embodiment, it is assumed that the compulsory force regarding work is stronger than the compulsory force regarding a sports gym. The strength of the coercive force depends on the penalties for the subject S for neglecting work or going to a sports gym or the like, or the evaluation of the subject S by a third party.

The constraint information may indicate, for example, the date and time when the subject S performed the activity, the restraint time by the activity, and the activity location. Note that the restraint time is not limited to the time from the start to the end of work or the like, and may include the time to move to arrive at the work place before the start of work or the like.
When the constraint information is represented by a date and time, the constraint information adding unit 12 sets, for example, activity information acquired a plurality of times from 0:00 to 24:00 as self-activity information for one day. The self activity information is associated with at least one of the year, month, day, and day of the acquisition date as constraint information for each day of self activity information (activity information unit described later). In such processing, the calculation unit 10 may include a calendar function for specifying a day and a clock function for acquiring time, and may add restriction information to the received self-activity information. In the first embodiment, the smartphone P may acquire the date and time by using a calendar function or a clock function, and transmit the date and time to the server device 1 together with the self-activity information. In such a case, the constraint information adding unit 12 stores the self activity information in association with the date and time when the self activity information is received. The “addition” of the constraint information referred to in the first embodiment means that the constraint information is stored as a set with the corresponding self-activity information, or is stored in association with the constraint information.

  The constraint information according to the first embodiment is not limited to the date and time described above. For example, the constraint information may be represented by a schedule of the subject S such as “working in Tokyo”, “business trip to Osaka”, “holiday”, and “use of a gym”. When the schedule is set as the constraint information, the subject S may register his / her own schedule in the server device 1 in advance, or the smartphone P may transmit the self-activity information to the server device 1 together with the schedule of the subject S. Is also good.

(Group creation department)
After the self-activity information is acquired over a plurality of days, the group creation unit 13 groups the self-activity information or a group of the self-activity information in descending order of similarity. At this time, in the first embodiment, a group of self-activity information is defined as a plurality of self-activity information acquired during a certain period of time as an activity information unit. Then, the group creation unit 13 classifies the activity information units based on the similarity of the tendency of the activity information unit to change over time in the activity amount. In the first embodiment, the “certain time” is set to 24 hours, and the self activity information acquired during one day is set as an activity information unit. The tendency of the activity amount to change over time refers to a tendency that the activity amount does not change over time, or increases or decreases. In addition, the similarity of the tendency refers to coincidence or disagreement of a time zone where the amount of activity does not change. The tendency similarity refers to coincidence or disagreement of time periods during which the amount of activity increases or decreases. Further, the similarity of the tendency indicates agreement or disagreement of the degree of increase or decrease in the amount of activity.
However, the first embodiment is not limited to using the self-activity information acquired during 24 hours as an activity information unit. The activity information unit may be a shorter time, for example, 6 hours or 12 hours, or may be a longer time, such as one week or one month.
In addition, the process of classifying the acquired single self-activity information includes, for example, acquiring an activity amount at a fixed time (for example, 3 o'clock, 9 o'clock, 15:00, etc.) for a plurality of subjects, and It is conceivable to group the closeness of the value of the activity amount acquired at the time as the similarity.

  Further, the group creation unit 13 of the first embodiment has an activity pattern extraction unit 14. The activity pattern extraction unit 14 executes an activity pattern extraction step of extracting a self activity pattern from self activity information. The activity pattern extraction unit step of the first embodiment, based on the self activity information included in at least one of the groups, generates a self activity pattern indicating a tendency of a temporal change in the activity amount of the subject S in units of an activity information unit. This is the extraction step. In addition, the self-activity pattern of the first embodiment is represented by any of a straight line, a curve, and a polygonal line representing a change in the amount of activity with time. In the first embodiment, a self activity pattern is extracted from the self activity information of each day, and the self activity information (activity information unit) is grouped according to the similarity of the shape of the self activity pattern.

FIGS. 3A to 3E are diagrams showing self-activity information and self-activity patterns extracted from the self-activity information. FIG. 3A is a diagram illustrating a state in which self-activity information of a 52-year-old male subject S is grouped. FIG. 3A is a table showing the acquisition time zone and the number of steps as self-activity information acquired from 0:00 to 24:00. One row of the table shows an activity information unit which is one day's own activity information, and FIG. 3A shows the number of steps for 28 days. The horizontal axis of FIG. 3A indicates a time zone in which the self-activity information is acquired every three hours. The vertical axis indicates the constraint information attached to each activity information unit. The self activity information shown in FIG. 3A is classified into a group α, a group β, a group γ, and a group δ according to the degree of similarity for each activity information unit. FIG. 3 (b) shows the self-activity pattern P _alpha extracted from self activity information belonging to the group alpha. The activity information units belonging to the group α since the plural, and the plurality of activity information unit self activity patterns P average of a group of self-activity pattern extracted from each _α of the first embodiment. FIG. 3 (c) shows the self-activity pattern P _beta extracted from activity information units belonging to the group beta. FIG. _3D shows a self-activity pattern _Pγ extracted from the activity information unit belonging to the group γ, and FIG. 3E shows a self-activity pattern _Pδ extracted from the activity information unit belonging to the group δ. I have.

  The creation of a group by the group creation unit 13 can be performed until all the activity information units finally belong to one group. However, in the first embodiment, the number of groups suitable for comparing and analyzing the activity state of the subject S is set in advance. In the first embodiment, the number of groups is four.

According to the self activity information shown in FIGS. 3A and 3B, it can be understood that the ratio of the self activity information acquired on Saturday and Sunday is large in the group α. Also, the self-activity patterns P _alpha group alpha, (be between the increasing and decreasing number of steps, the most significant value in this period) the peak value of the number of steps during the day than the self activity pattern is extracted from the other groups Is also small. For this reason, in the first embodiment, it is estimated that the activity amount of the subject S, who is a generation in the occupational area, decreases because he or she does not walk on a holiday than on weekdays. In the first embodiment, the group α in which the ratio of the self-activity information acquired on Saturday and Sunday is large can be, for example, a “holiday” group.

According to the self-activity information shown in FIGS. 3A and 3C, the group β belongs to the least number of activity information units in the group, and the number of steps indicated by the self-activity pattern P _β is 12 It has increased since the hour and reached a peak from 24:00 toward the next day. In the first embodiment, the group β having a large amount of activity acquired at midnight can be a group of “scheduled after employment”, for example. Further, according to the self activity patterns shown in FIGS. 3A and 3D, the group γ has the largest number of activity information units belonging to the group, and the self information acquired on all days from Monday to Friday. Contains activity information. The self-activity patterns _P γ of the group γ, the number of steps is increased to early morning from midnight, it has been shown to reduce once in the morning. Also, the self-activity patterns P _gamma group gamma, 18 step count to 21 o'clock is increased again from time, have been shown to reduce the 21 o'clock.

Further, according to the self-activity information shown in FIGS. 3A and 3E, it can be understood that the group δ is a group to which the self-activity information mainly acquired on Wednesday and Tuesday belongs. The Group self activity pattern P _[delta] of [delta], has once step count peaks between 6 o'clock and 9 o'clock, 18 step count to 21 o'clock is increased again from time, it illustrates a reduction of 21 o'clock ing. In the first embodiment, the group γ and the group δ having a large activity amount acquired in the morning and from 18:00 to 21:00 can be, for example, a “normal work” group.

From the self-activity information described above, the first embodiment estimates that the subject S is active according to a plurality of self-activity patterns even on weekdays. Further, in the first embodiment, since the self-activity information belonging to one group has bias in day of acquisition date, a plurality of subjects S is according to days of the week from the self-activity patterns P _alpha self activity pattern P _[delta] Presumed to have. Further, in the first embodiment, is between from the fact that the tendency of between 18 o'clock in the self-activity patterns P _gamma and self activity pattern P _[delta] until 21 similar, to 21 o'clock 18 step count Is estimated to be due to the activity related to the return of the subject S.

  In the first embodiment, the group creation unit 13 makes a determination such as “the ratio of the self activity information acquired on Saturday and Sunday is large in the group α” by the statistical processing. The group creation unit 13 determines whether the difference between the days of the week when the activity information units included in the group are acquired by the statistical processing is significant or accidental. In the first embodiment, the significance of the large number of activity information units on Saturday and Sunday included in the group α is, for example, the number of activity information units on Saturday and Sunday between the group α and another group. Performed by comparison. At this time, in the first embodiment, the probability that the difference in the number of activity information units on Saturday and Sunday obtained by comparison is obtained by chance is calculated by the chi-square test. If the probability obtained as a result of the calculation is smaller than a predetermined significance level (for example, 0.05), the group creating unit 13 determines that the group α includes many activity information units on Saturday and Sunday. I do. It should be noted that such a determination is, of course, preferably performed after a sufficient number of pieces of self-activity information have been accumulated to obtain process reliability.

  FIG. 4A is a diagram illustrating a state in which self-activity information targeting a 76-year-old man is grouped for comparison with a 52-year-old man subject S. FIG. 4A is a table showing the acquisition time zone and the number of steps as self-activity information acquired from 0:00 to 24:00. One row of the table shows one day of self activity information, that is, activity information units, and FIG. 4A shows 28 days, that is, the number of steps in the activity information units for 28 days. The horizontal axis of the table shows the time zone of self-activity every three hours. The vertical axis indicates the constraint information attached to each activity information unit.

The self activity information shown in FIG. 4A is classified into a group ε, a group ζ, and a group η according to the similarity. FIG. 4B shows the self activity pattern _Pε extracted from the activity information units belonging to the group ε. FIG. 4C shows a self-activity pattern P _# extracted from the activity information units belonging to the group β. FIG. 4D shows the self activity pattern P _η extracted from the activity information units belonging to the group γ. According to FIG. 4A, it can be seen that the self activity information belonging to the group ε is acquired on all days except Friday. According to FIG. 4B, the self-activity information belonging to the group I is acquired on the day of the week excluding Monday, Friday and Sunday. Most of the self-activity information belonging to the group η is acquired on Friday.

  From the above, when the target person is out of the occupational generation, although he / she is active according to a plurality of self-activity patterns, compared to the target generation S of the occupational generation, the day of the week on which the self-activity information is obtained in one group is different. It can be seen that the bias is small. This indicates that the subjects who are out of the occupational generation have weaker social constraints than the subjects S of the occupational generation, and tend to be active without depending on the day of the week. As described above, in the first embodiment, the number of groups created by the group creation unit 13 is set in advance, so that when the self-activity information of the subject having a weak social constraint is classified, it is classified into one group. The dispersion of the self-activity information to which it belongs becomes relatively large.

As described above, in the first embodiment, since the activities of the subject S are classified according to their similarities, it is possible to create a group in which the self-activity information has little variation. In the first embodiment, by adding the constraint information to the created group, the self-activity information and the self-activity patterns of the groups to which the same constraint information has been added can be compared. At this time, in the first embodiment, the acquisition time may be different for one subject S, and the self-activity information or the self-activity pattern of the group having the same constraint information may be compared. The self-activity information and self-activity patterns of the groups may be compared. Comparing self-activity information or the like of a group of one subject S whose acquisition time is different and whose constraint information is the same is useful for detecting a change in activity of the subject S due to aging or the like. In addition, comparing the self-activity information and the like of the same group with the same constraint information for different subjects can provide the subject with the opportunity to objectively review his / her own activity status by comparing it with others.
In addition, the first embodiment can suppress variation in self-activity information belonging to one group by applying to a job-occupation generation with a strong social constraint. Therefore, the first embodiment is suitable for classifying the self-activity information of the job generation.

(Mental and Physical Information Creation Department)
The body and mind information creation unit 15 creates mind and body information on the body or spirit of the subject S based on self-activity information included in the group. The above operation is an information creation step in the first embodiment. In the first embodiment, the mental and physical information is converted by the output unit 16 into transmission data that can be transmitted by the transmission device 5. The transmission data is transmitted from the transmission device 5 to the smartphone P and displayed on the subject S.

For example, the mind and body information creating unit 15 determines the number of steps of the subject S from the self activity information (self activity pattern) of the subject S. Then, when the total number of steps per day of the subject S is less than a preset reference value, mind and body information for recommending walking is created.
In the first embodiment, a message based on mind and body information may be created. For example, when the mental and physical information creating unit 15 determines from the self-activity information of the subject S that the number of steps of walking performed in the morning and evening is small, it creates a message such as "Why not get off in front of one station and walk?" You may. Further, when the total number of steps in the day is less than the reference value and the number of steps in the morning and evening is equal to or greater than the reference value, the mind and body information creating unit 15 outputs a message such as "How about walking to the park at noon?" May be created. The server device 1 can acquire information on the position of the target person S and the weather in the area where the target person S is located. From this, the body and mind information creation unit 15 can create a message such as "Today is good weather, so let's walk to the shrine to see the cherry blossoms."

  Further, in creating the message, in the first embodiment, other information may be used together with the self-activity information. As other information, for example, health information on the health of the subject S or schedule information on a schedule can be considered. When using the health information of the subject S, the server device 1 acquires health information on the body and spirit of the subject S, and further on lifestyle. As the health information, for example, the age, gender, pulse, blood pressure, height, weight, sleep time, presence or absence of smoking and drinking of the subject S, subjective emotions of the subject S (motivated, tired) Etc. Such information may be information that the subject S continuously transmits to the server device 1 and makes a self-application. In the first embodiment, the blood pressure, the heart rate, the blood oxygen, the visual acuity, the respiration rate, the hearing, and the lung capacity of the subject S acquired by the application of the smartphone P are automatically transmitted from the smartphone P to the server device 1. You may do so. Further, in the first embodiment, the server device 1 may acquire the result of the health examination of the subject S from the medical institution and store the result in the database 7.

  When using the health information, the mind and body information creating unit 15 extracts items to be improved for the subject S from the health information, for example. When the item to be improved is excessive weight or high blood pressure, the mental and physical information creating unit 15 determines the number of steps of the subject S in the same manner as described above. If the total number of steps per day of the subject S is less than a preset reference value, for example, a message such as "If you walk for 15 minutes every day, you can expect a weight loss effect of 2 kg in three months" Can be created.

  In the case where the schedule information is used, in the first embodiment, the subject S may transmit schedule information including information on his / her own work day or holiday through the smartphone P in advance, for example. The transmission of the schedule information may be such that the subject S inputs the schedule information on an input screen provided from the server device 1 side, for example. Further, the transmission of the schedule information may be, for example, a method in which the smartphone P extracts the schedule information from the information of the target person S recorded on the smartphone P, and transmits and registers the schedule information to the server device 1. The mind and body information creating unit 15 can provide a message to the subject S based on the mind and body information and the schedule information of the subject S.

  When the schedule information is used, for example, when the subject S indicating that the schedule information is a holiday does not substantially walk until afternoon, it can be estimated that the subject S is sleeping until the afternoon. At this time, the body and mind information creation unit 15 can create a message such as "If the rhythm of life is disrupted, the work efficiency of weekdays will decrease, let's wake up soon." Further, for example, the mental and physical information creation unit 15 creates a message such as "Thank you for your late work, let's go home without detouring for tomorrow" for the subject S indicating that the schedule information indicates that the user is on a business trip. can do.

  The mind and body information creating unit 15 executes an information creating step and creates mind and body information or a message including mind and body information based on the tendency of the constraint information attached to the self-activity information included in each group. At this time, the information creation step further includes a constraint tendency acquisition step of acquiring a tendency of the constraint information, and creates psychosomatic information on the subject's body or spirit based on the acquired tendency of the constraint information. The “trend of the constraint information” indicates, for example, that when the constraint information includes a day of the week, the acquisition date of the self-activity information belonging to each group has many Saturdays or no Wednesday. Specifically, for example, when creating a group, the group creating unit 13 determines the similarity between the newly acquired self-activity information and the previously acquired self-activity information. At this time, the group creation unit 13 compares the self-activity information acquired today (Saturday) with the self-activity information of a group including many Saturdays (a holiday group) as constraint information. Then, when the group creating unit 13 determines that the similarity between the two is lower than the preset degree, the group creating unit 13 may notify the mental and physical information creating unit 15 of the determination result. For example, if the self-activity information acquired today indicates that the subject S has been active earlier in the morning than on a Saturday in the past, the mental and physical information creating unit 15 determines that the subject S is We recommend that you take a nap before. "

  The determination of the similarity between the self-activity information acquired today and the self-activity information acquired in the past can be performed at an arbitrary timing. For example, the determination of the similarity between the two may be performed at the timing when the amount of activity for one day is acquired, or may be performed at 12:00 and 24:00. Further, the mental and physical information creating unit 15 can change and adjust the message according to the acquisition timing of the self activity information having low similarity.

  FIG. 5 illustrates a state in which the message created by the mind and body information creating unit 15 is transmitted from the transmitting device 5 and displayed on the smartphone P. In the example illustrated in FIG. 5, the message M1 of “It is time to get tired due to the rhythm that has collapsed on holidays. Let's relax” is displayed on the display screen 20 of the smartphone P. Such a message M1 is a message when self-activity information having low similarity to self-activity information acquired in the past is acquired at the beginning of the week. The first embodiment is not limited to displaying the message M1 on the display screen 20. For example, the message M1 may be notified to the target person S by voice, or may be transmitted to another communication device registered in advance and displayed.

  Further, in the first embodiment, in the information creating step, information indicating an article or agent recommended to be used by the subject S may be created. When creating information indicating an article or an agent, the server device 1 may, for example, encourage walking when commuting or at lunch, encourage walking on holidays, recommend rest, recommend activity start or end time change, and sleep time. An article or agent is determined in advance for each piece of psychosomatic information such as an indication, and information on the article or agent is stored in the database 7 together with the mind and body information. Then, the mind and body information creating unit 15 creates a message including information recommending an article or an agent corresponding to the mind and body information. In the example shown in FIG. 5, the mind and body information creation unit 15 creates a message M2 “Let's heal the tiredness of the day in a warm bath. Bath salts are recommended”. The message M2 is displayed on the display screen 20 of the smartphone P together with the recommended image of the agent d.

  The agent referred to in the first embodiment includes so-called personal care products for the purpose of cleaning, body taste, taste, and the like, excluding pharmaceuticals, drugs, supplements, and the like. Examples of personal care products include face care products, body care products, oral care products, in-bath health care products, and hair care products. The articles in the first embodiment include a supporter, a sphygmomanometer, a thermometer, a disposable diaper, a mask, a heating article, and the like. Among the above-mentioned products, for example, when the blood pressure of the subject S is higher than a preset reference value or when the blood pressure of the subject S tends to increase over time, it is recommended to use a supplement that is effective for suppressing blood pressure increase together with walking. Is done. It is also recommended to use a sphygmomanometer to measure and manage blood pressure daily. In addition, for example, on a day when it is determined that the working time is particularly long based on the peak interval of the number of steps related to commuting of the subject S, the mental and physical information creating unit 15 recommends a rest and recommends a heating article that warms the eyes. May be created.

  FIG. 6 is a flowchart for explaining the activity classification method executed by the server device 1 which is the activity classification device of the first embodiment described above. In the first embodiment, when the subject S receives the service of the server device 1, first, his / her schedule information is input to the server device 1. In the first embodiment, the subject S registers health information such as name, age, height, and weight together with the schedule information. At this time, in the first embodiment, the number N of groups created by the group creation unit 13 is input to the input unit 11 by the operator of the server device 1 or a service start program (step S101). The input of the schedule information may be performed by the subject S transmitting the schedule information or the like from the smartphone P or another communication device, for example. In addition, the registration of the health information may be such that the target person S applies for the provision of a service by a document or the like, and the operator of the server device 1 inputs the information to the input unit 11.

  Next, the server device 1 receives information indicating the activity amount of the target person from the smartphone possessed by the target person by the receiving device 3 at predetermined time intervals. The input unit 11 generates self-activity information of the target person by associating the activity amount with a reception time zone of the information indicating the activity amount. Through such processing, the arithmetic unit 10 acquires the activity information (Step S102). The constraint information adding unit 12 adds constraint information to the acquired self-activity information (Step S103). To add the constraint information, for example, the constraint information adding unit 12 extracts the information of the work day and the holiday from the schedule information, compares the information with the reception date and time of the activity information and the day of the week, and adds “work day” or “holiday” to the activity information for each day. "May be added.

Next, after the self-activity information is acquired for a predetermined number of days (for example, for one month), the group creation unit 13 determines the similarity of the self-activity information for each activity information unit and divides the self-activity information into a plurality of pieces. Classify into groups. Such processing is realized by a clustering method. In the first embodiment, a group is created by, for example, an agglomerative clustering method using a hierarchical method. In the aggregation type clustering, an initial state is created in which a plurality of clusters C _m each including one activity information unit x _m exist, and a distance (dissimilarity) D (x 1, x 2) between the activity information units x 1 and x 2 is created. ), The distance D (C1, C2) between the clusters is calculated (step S104). Then, the group creation unit is performed by sequentially merging the clusters in ascending order of the distance D (C1, C2) between the clusters (step S105). The creation of a group is repeatedly executed each time self activity information is acquired on a new day. In the first embodiment, clustering is repeated until the number of groups (clusters) reaches the number N of groups input in step S101.

In the first embodiment, it is preferable that the latest self-activity information grouped is periodically stored in the database 7 for long-term storage. The long-term storage of self-activity information is preferably, for example, about six months to twenty years. In this way, in the subsequent message creation (information creation step), mental and physical information may be created for one subject S by comparing at least two pieces of self-activity information having different acquisition times. According to such processing, it becomes possible to analyze the self activity information and analyze the relationship between the exercise state of the subject S and the age, the illness, and the environment.
More specifically, for example, the body and mind information creation unit 15 obtains the self-activity information of the subject S obtained on a holiday at time T1 and the self-activity information obtained on a holiday of time T2 after the time T1. To create mental and physical information. Specifically, when the activity amount acquired at the time T2 is larger than the activity amount acquired at the time T1, the mental and physical information creating unit 15 compares the health information of the subject S between the time T1 and the time T2. It is possible to do. If the weight of the subject S has decreased at the time T2 compared to the time T1, for example, a message such as "The effect of the exercise on holidays has appeared. Let's work hard to the target weight in this condition" is created. You may.

  Next, the activity pattern extraction unit 14 of the group creation unit 13 extracts a self activity pattern for each group created by the group creation unit 13 (Step S106). In the first embodiment, a self-activity pattern may be periodically stored in the database 7 together with the self-activity information described above.

  Next, the mind and body information creating unit 15 determines whether or not it is necessary to output mind and body information from the self activity information of the subject S (step S107). The determination of the necessity of the mental and physical information output is performed, for example, when the minimum value of the dissimilarity between the activity information units of the subject S is larger than a preset reference value in step S104. You may. In the first embodiment, when encouraging the target S to perform an activity, a message including mental and physical information may be output to the target S at a timing suitable for the activity (step S107: Yes). When the mental and physical information is not output (Step S107: No), the server device 1 repeatedly acquires the self activity information.

When outputting the body and mind information (Step S107: Yes), the mind and body information creation unit 15 considers the output timing, the situation, the position, the climate, the state of the activity on the previous day, and the like of the subject S, and the message including the mind and body information. Create The creation of the message is performed, for example, by the mind and body information creating unit 15 selecting a template created in advance corresponding to the mind and body information. The selection of the template may be performed in accordance with the situation of the target person S or the like, or may be performed using a random number in order to prevent the same template from being repeatedly selected in a short cycle. Further, the template may be stored in the database 7.
The mind and body information creating unit 15 determines the date and time, place, name, and the like of the selected template according to the subject S. Further, the information about the agent stored in the database 7 or the like is read out in correspondence with the mental and physical information to complete the message. The completed message is output from the output unit 16 to the transmitting device 5 (step S108), and is transmitted to the smartphone P or the like owned by the subject S.

  In the first embodiment described above, the activity information of the subject S is grouped according to the similarity. When the target person S is a generation in the work area, it is considered that the target person S goes to the same workplace at substantially the same time and leaves the workplace at substantially the same time and returns home. Under such social restrictions, the activity information belonging to one group has little variation. The first embodiment can detect a change in the activity of the target person S with high accuracy by detecting the change in activity information with little variation.

In the first embodiment, for example, activity information on a plurality of days is divided into groups by paying attention to similarities without being conscious of social restrictions. For this reason, the activity information of the first embodiment creates a group that reflects repetitive business trips, overtime, and post-employment activities even on the same weekday, and accurately grasps the status of the activity of the target person S. be able to.
Further, the first embodiment is not limited to notifying the subject S of the physical and mental information. In the first embodiment, for example, mental and physical information may be notified to the workplace, family, or hospital of the target person S.

[Second embodiment]
Next, an activity classification method and an activity classification device according to a second embodiment of the present invention will be described. The activity classification device according to the second embodiment has the same configuration as the activity classification device according to the first embodiment shown in FIG. Therefore, in the second embodiment, illustration of the activity classification device is omitted, and a part of the description is omitted.
The second embodiment aims to bring self-activity information closer to comparative activity information by comparing self-activity information with activity information (comparative activity information) of another target person (hereinafter, referred to as “others”). It is. Therefore, in the second embodiment, as in the first embodiment, the server device 1 acquires the self-activity information of the subject S, and classifies the information into a plurality of groups according to the degree of similarity to each other. In parallel with this, the server device 1 acquires self-activity information from a plurality of subjects, respectively, and classifies them into a plurality of groups according to their similarity. In the second embodiment, information on the body and health of the subject S and a plurality of other subjects is stored in the database 7.

  In the second embodiment, the information creation step creates the mind and body information by comparing the index designation step of designating an index relating to the body or performance with the comparison activity information of the other person corresponding to the designated index and the self activity information. And generating comparison information. Here, as described above, the “other person” may be any person as long as it is a target person for acquiring activity information. The index relating to the body may be, for example, a target value such as body fat, muscle mass, chest circumference, and abdominal circumference. The “other corresponding to the index” refers to a person having a designated index, for example, a specific body fat or the like, among the above-mentioned others. Further, the index relating to the body may not be the target value itself, but may be one that designates the designation of another person having the target body of the target person S.

  The performance in the second embodiment indicates physical ability, health or mental state. The physical ability can be designated, for example, as mountain climbing Fuji, crawl 100 m, or the like. In the second embodiment, such an index can be set together with the age and gender of a 50-year-old man and a 40-year-old woman. Further, as the health condition, a problem to be solved by the subject S such as no back pain and no stiff shoulder can be set. Further, it is possible to set a target health condition in the future such as walking without a cane at the age of 75 in the future and no dementia at the age of 85 in the future. The mental state may be a state obtained by self-report of another person, such as absence of insomnia and motivation to work.

FIGS. 7A to 7F are diagrams showing comparison activity information of a plurality of other persons grouped, and each of the patterns (comparison) extracted after averaging the plurality of comparison activity information. Activity pattern). 7A to 7F, the vertical axis represents the number of steps, and the horizontal axis represents the time period for acquiring the number of steps. Comparative activity pattern P _theta shown in FIG. 7 (a), after having a peak in the morning, has risen in the afternoon toward 24:00. The comparative activity pattern P _ι shown in FIG. 7B has a peak in the morning and then peaks in the afternoon and falls at 24:00. The morning and afternoon peaks are about the same height. The comparative activity pattern P _κ shown in FIG. 7C has a peak in the morning and then falls at 24:00 with a peak in the afternoon. The morning peak is higher than the afternoon peak. The comparative activity pattern P _λ shown in FIG. 7D has a peak in the morning and then peaks in the afternoon and falls at 24:00. The morning peak and the afternoon peak are approximately the same height, but the morning peak has a certain number of steps regardless of time. Comparative activity pattern P _mu shown in FIG. 7 (e) of the comparative activity pattern shown in FIG. 7 shows the least number of steps. The comparative activity pattern _Pμ has a relatively low peak in the morning, the period including noon, and the afternoon, and the peak in the morning and the peak in the afternoon are approximately the same height. Comparative activity pattern P _mu shown in FIG. 7 (f) of comparative activity pattern shown in FIG. 7 shows the least number of steps. The comparative activity pattern _Pv has relatively low peaks in the morning and noon to night. The peak in the morning is slightly lower than the peak in the afternoon and night.

In the second embodiment, for example, when the subject S sets “body fat 15% or less” as a physical index, the mental and physical information creating unit 15 extracts the comparative activity pattern _Pν from the comparative activity pattern _Pθ . Among the groups, a group including many comparison subjects whose body fat is 15% or less is specified by statistical processing from health information. Then, the comparison activity pattern extracted from the specified group is compared with the self activity information of the subject S, and a message for making the self activity pattern extracted from the self activity information close to the comparison activity pattern is created. Such a message may be, for example, "Let's go to the gym twice a month on holidays, and pay attention to meals".

Further, in the second embodiment, for example, when the subject S sets “walking without a cane at the age of 75 in the future” as an index of the health condition, the mental and physical information creating unit 15 can walk without using the cane from the health information. Elderly Elderly as Others Then, from the comparative activity information acquired and grouped in the past, the group including the selected subject most is specified by statistical processing. Next, in the second embodiment, the comparative activity information extracted from the specified group is compared with the self activity information of the subject S, and the self activity pattern extracted from the self activity information is approximated to the comparative activity pattern. Create a message. Such a message may be, for example, "Let's walk little by little every day, and refrain from using the escalator when commuting".
In the second embodiment described above, when the activity classification method of the second embodiment is applied to a subject Q other than the subject S, the self activity information acquired from the subject S is based on the self activity of the subject Q. Treated as comparative activity information that is compared with the information.

  According to the second embodiment described above, by inputting the target index, the target person S can be active in the same manner as the target other person. From this, according to the second embodiment, it is possible to suggest a method of activity that is advantageous for achieving the target for the target person S.

DESCRIPTION OF SYMBOLS 1 ... Server apparatus 3 ... Receiving apparatus 5 ... Transmitting apparatus 7 ... Database 10 ... Operation part 11 ... Input part 12 ... Restriction information addition part 13 ... Group creation part 14 ... activity pattern extraction unit 15 ... mind and body information creation unit 16 ... output unit 20 ... display screen

Claims (11)

An information acquisition step of acquiring self activity information indicating the amount of activity of the target person for each time period a plurality of times for the target person,
A plurality of self-activity information, or a group of the self-activity information, a group creation step of creating a plurality of groups by classifying according to their similarity,
An activity classification method, including:   The group of the self activity information is an activity information unit that is a plurality of the self activity information acquired at a certain time, and the group creation step includes a trend of a temporal change of the activity amount in the activity information unit. The activity classification method according to claim 1, wherein the activity information units are classified according to the similarity of the activities.   An activity pattern extracting step of extracting, from the self activity information included in at least one of the groups, a self activity pattern indicating a tendency of the activity amount to change with time in units of the activity information unit, further comprising: 2. The activity classification method according to 2.   The activity classification method according to any one of claims 1 to 3, further comprising an information creating step of creating psychosomatic information on the subject's body or mind based on the self-activity information included in the group.   The activity classification according to any one of claims 1 to 4, further comprising a constraint information adding step of associating the self-activity information with constraint information relating to a constraint on the behavior of the subject at the time of acquiring the self-activity information. Method.   The activity classification method according to claim 4, wherein in the information creating step, a message to the subject is provided by health information on the health of the subject or schedule information on a schedule of the subject along with the self-activity information.   The information creating step is an index specifying step of specifying an index related to body or performance, and comparing the self-activity information with comparison activity information that is activity information of another person corresponding to the specified index, the mental and physical information. The activity classification method according to claim 4, further comprising: a comparison information generating step of generating   The activity classification method according to claim 4, wherein in the information creating step, the mental and physical information is created by comparing at least two pieces of the self activity information having different acquisition times for the subject.   The activity classification method according to claim 4, wherein in the information creating step, information on an article or an agent recommended to be used for the subject is created.   The information creating step further includes a constraint tendency acquiring step of acquiring a tendency of constraint information relating to a constraint on the behavior of the subject at the time of acquiring the self-activity information, based on the tendency acquired by the constraint tendency acquiring step. The activity classification method according to claim 4, wherein mental and physical information relating to a body or a psyche of the subject is created. An information acquisition unit that acquires self activity information indicating the amount of activity of the subject for each time zone a plurality of times for the subject,
A plurality of the self-activity information, or a group of the self-activity information, a group creating unit that creates a plurality of groups by classifying according to the degree of similarity to each other
An activity classification device comprising:

Images