Program, information processing method, information processing device, and learning model generation method

Abstract

To provide a program, etc. that monitors the motivation of an owner by utilizing an individual person's smartphone.SOLUTION: A program makes a computer 20 execute processing that acquires feature quantity data obtained through the use of a smartphone 30 by a user, and outputs engagement information by inputting the acquired feature quantity data into an engagement model 46 that outputs engagement information related to engagement when feature quantity data is inputted.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a program, an information processing method, an information processing apparatus, and a learning model generation method.

Background Art

[0002] There has been proposed an information presentation system that analyzes data collected using a device incorporating a sensor, such as a smartwatch, and presents sleep advice and activity advice (Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] It is socially desirable to detect at an early stage a person showing a decline in motivation and provide support such as counseling. The information presentation system of Patent Document 1 can also be used for detecting a person in need of support.

[0005] However, in order to use the information presentation system of Patent Document 1, each person needs to always wear a device such as a smartwatch having sufficient functions and performance. For a person who has difficulty always wearing a device, the information presentation system of Patent Document 1 cannot be used.

[0006] On the other hand, in modern society, a smartphone is an essential tool in daily life. Many people carry and use a smartphone on a daily basis.

[0007] On one aspect, it aims to provide a program or the like that monitors the motivation of the owner using the smartphone owned by an individual person.

Means for Solving the Problem

[0008] The program acquires feature amount data obtained with the use of a smartphone by a user, and causes a computer to execute a process of outputting engagement information by inputting the acquired feature amount data into an engagement model that outputs engagement information regarding engagement when the feature amount data is input.

Advantages of the Invention

[0009] In one aspect, it is possible to provide a program or the like that monitors the motivation of an individual by using the individual's smartphone.

Brief Description of the Drawings

[0010]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Modes for Carrying Out the Invention

[0011] [Embodiment 1] FIG. 1 is an explanatory diagram for explaining the information processing system 10 according to Embodiment 1. In the present embodiment, an information processing system 10 will be described that monitors the employment motivation of university students and notifies supporters such as school counselors and the students themselves of information regarding students whose employment motivation has declined.

[0012] There are students who enter university but whose employment motivation declines after enrollment, resulting in truancy and ultimately dropping out. For example, students whose employment motivation has declined due to stress caused by a change in environment or disruption of their lifestyle habits, etc., are likely to regain their employment motivation if supporters such as school counselors intervene and provide support at an early stage.

[0013] However, with the spread of distance learning, it has become difficult for people around the student, such as classmates, seniors, and faculty members, to notice changes in the student's condition at the stage before truancy. In distance learning, it is difficult to contact students who have become truant and are no longer connected to the university network, so it is also difficult for support to reach them. Therefore, it is important to discover early students whose employment motivation has declined and start appropriate support.

[0014] The information presentation system disclosed in the aforementioned Patent Document 1 can also be used to discover students in need of support. However, in order to use the information presentation system of Patent Document 1, each student needs to purchase a device such as a smartwatch with sufficient functions and performance, and constantly wear it while regularly charging it. For students who have difficulty purchasing and properly wearing the device, the information presentation system of Patent Document 1 cannot be used.

[0015] On the other hand, for the student generation, smartphones are essential tools in daily life. Many students carry and use smartphones on a daily basis. In the present embodiment, data is collected from individual student smartphones to monitor the employment intention.

[0016] Using FIG. 1, the outline of the information processing system 10 for monitoring the employment intention of students will be described in three stages: the data collection stage, the engagement model generation stage, and the utilization stage.

[0017] [Data Collection Stage] In the data collection stage, raw data is collected from the smartphones 30 of the students to be monitored and stored in the server 20. It is desirable that the students to be monitored are selected so that the distribution of age, gender, department affiliation, etc. is close to the distribution of the actual enrolled students.

[0018] From each smartphone 30, the raw data collected by a sensor 37 (see FIG. 3) or the like built into the smartphone 30 is transmitted to the server 20. The transmission of the raw data is appropriately carried out at a timing that does not affect the use of the smartphone 30, such as when the student is not operating the smartphone 30. The transmission of the raw data may be carried out at a timing that does not incur communication costs, such as when connected to a wireless LAN (Local Area Network).

[0019] Questionnaires are displayed on the smartphone 30 at appropriate times such as in the morning and at night. The answers entered by the students targeted for monitoring by operating the smartphone 30 are sent to the server 20. The questionnaire includes questions regarding employment motivation.

[0020] In this embodiment, the questions regarding employment motivation are the questions used in UWES (Utrecht Work Engagement Scale), modified for students. UWES is used in engagement surveys targeting employees of companies. Engagement means the attachment, dedication, enthusiasm and passion of employees towards their work, as well as the desire for self-growth through work, etc.

[0021] In UWES, the engagement that employees have towards "work" itself is scored on three scales: "Vigor", "Dedication" and "Absorption". "Vigor" is a scale indicating the energy and psychological resilience during employment. "Dedication" is a scale indicating the degree of involvement in work, the degree of meaningfulness regarding work, and the degree of pride in work. "Absorption" is a scale indicating the degree of concentration and immersion in work.

[0022] There are three types in UWES: UWES-17 which uses a 17-item questionnaire, UWES-9 which uses a 9-item questionnaire, and UWES-3 which uses a 3-item questionnaire. Employees answer each item on a seven-point scale from "Never" to "Always feel". Based on the answer results, scores regarding each of the above three scales are calculated.

[0023] In this embodiment, the UWES-3 questionnaire was modified for students and used. The questionnaire has a total of three questions, one question regarding each of the scales of "Vigor", "Dedication" and "Absorption". Examples of the modified questions will be described later.

[0024] The students to be monitored answer each question on a seven-point scale from "never" to "always feel". "Never" is converted to a score of "0 points", and "always feel" is converted to a score of "6 points", which is used for engagement information regarding each measure of "vitality", "enthusiasm", and "dedication". Based on the answers to each of the three items, a score regarding the comprehensive engagement information of the students is calculated.

[0025] The comprehensive engagement information is, for example, the average value of the engagement information of the three items. The comprehensive engagement information may be the total value of the engagement information of the three items. The calculation result such as the value obtained by adding a predetermined weight to each of the three items may be used as the comprehensive engagement information. Any one or two of the three items may be used as the comprehensive engagement information. In the following description, the comprehensive engagement information may sometimes be simply described as engagement information.

[0026] Note that the questionnaire used for the evaluation of employment intention is not limited to the modified version of UWES-3. The modified version of UWES-9 or UWES-17 may be used. The evaluation scale of employment intention is not limited to UWES. For example, any scale such as personal engagement focusing on the role at work and organizational commitment focusing on the strength of the connection between employees and the workplace organization, which is modified for the evaluation of employment intention, may be used. A new scale may be developed and used to be suitable for the evaluation of employment intention.

[0027] The control unit 21 (see FIG. 3) divides the raw data transmitted from the smartphone 30 into unit times such as 30 minutes or one hour, performs noise removal and statistical processing, etc., and calculates a representative value per unit time. The representative value is, for example, the arithmetic mean, geometric mean, maximum value, minimum value, mode, variance, or standard deviation. The control unit 21 may calculate a plurality of representative values for one item.

[0028] The control unit 21 creates feature amount data, which is time-series data arranging representative values in each unit time in time-series order. Thus, the control unit 21 converts the raw data transmitted from the smartphone 30 into feature amount data indicating the usage status of the smartphone 30. An example of items of the feature amount data is shown in Table 1.

[0029]

Table 1

[0030] In Table 1, GPS (Global Positioning System) means the position of the smartphone 30 measured based on radio waves of positioning satellites and radio waves of base stations, etc. Note that, in this embodiment, it is described as GPS, but the positioning satellite is not limited to the GPS satellite of the United States. For example, any positioning satellite such as the quasi-zenith satellite Michibiki can be used.

[0031] Note that the feature amount data shown in Table 1 is an example. Any feature amount data that can be created based on the raw data that can be acquired from the smartphone 30 may be created. Only a part of the feature amount data shown in Table 1 may be created.

[0032] The control unit 31 (see FIG. 3) of the smartphone 30 may create the feature amount data and transmit it to the server 20. Although the calculation load of the smartphone 30 is high, an information processing system 10 with a small communication volume can be realized. An information processing system 10 that can easily protect the personal information of students can be provided as compared with the case of directly sending the raw data to the network.

[0033] [Engagement model generation stage] The control unit 21 uses, as explanatory variables, feature quantity data for the day when the student to be monitored answered the questionnaire or for a predetermined period such as the past three days including the answer date, and uses engagement information based on the questionnaire as the objective variable to generate an engagement model 46 by machine learning. As the machine learning algorithm, a supervised machine learning algorithm for a known classification model such as random forest, CNN (Convolutional Neural Network), or transformer is used. The configuration of the engagement model 46 will be described later.

[0034] [Utilization stage] The program of the present embodiment is installed in the smartphone 30 that the student uses daily. Raw data is transmitted from each smartphone 30 to the server 20. The transmission of the raw data is appropriately performed at a timing that does not affect the use of the smartphone 30, such as when the student is not operating the smartphone 30.

[0035] The control unit 21 performs the same processing as in the above-described data collection stage on the raw data transmitted from the smartphone 30 and converts it into feature quantity data. The control unit 21 inputs the feature quantity data into the engagement model 46 to obtain engagement information.

[0036] The control unit 21 records the engagement information in association with, for example, the ID of each student. The control unit 21 extracts students with low engagement in employment, that is, students with low employment motivation, and notifies supporters such as school counselors and the students themselves. The supporter confirms the notification using an information processing device 50 such as a personal computer. The student himself / herself confirms the notification using his / her own smartphone 30.

[0037] Based on the time-series changes in the engagement information, the control unit 21 may extract students whose engagement has recently decreased and notify the supporters and the students themselves. Note that the criteria for extraction and the methods of notification may be different for the notification to the supporters and the notification to the students themselves. Also, it may be notified only to the supporters or only to the students themselves.

[0038] The control unit 21 may create a list of students determined to have low engagement and record it in a place easily accessible to the supporters. The supporters may provide support such as interviews with the students themselves and information exchange with the guardians as needed. Also, the control unit 21 may create a list of information about the students so that the supporters can view it as appropriate using the information processing device 50. From the perspective of personal information protection, it is desirable to grant the access right of the supporters only to the data related to the students who are the support targets.

[0039] The control unit 21 may automatically send messages such as reminders to students who meet certain conditions. The control unit 21 may automatically send messages such as situation confirmation to students who meet certain conditions and notify the supporters of the reply results.

[0040] FIG. 2 is an explanatory diagram for explaining the configuration of the engagement model 46. The engagement model 46 includes a vitality engagement model 461, an enthusiasm engagement model 462, and a dedication engagement model 463. The vitality engagement model 461 is a learning model that receives the input of explanatory variables selected from the feature data and outputs vitality engagement information. Here, the vitality engagement information regarding vitality is information that predicts the students' answers to the questions regarding vitality used in the data collection stage described with reference to FIG. 1.

[0041] Similarly, the enthusiasm engagement model 462 is a learning model that receives an input of explanatory variables selected from feature data and outputs enthusiasm engagement information. The immersion engagement model 463 is a learning model that receives an input of explanatory variables selected from feature data and outputs immersion engagement information. The enthusiasm engagement information and the immersion engagement information are also information for predicting students' answers.

[0042] The items of explanatory variables input to the vitality engagement model 461, the enthusiasm engagement model 462, and the immersion engagement model 463 may or may not be the same.

[0043] The engagement model 46 outputs engagement information, which is the average value of the vitality engagement information, the enthusiasm engagement information, and the immersion engagement information. The engagement model 46 may be configured to output the vitality engagement information, the enthusiasm engagement information, and the immersion engagement information in addition to the engagement information.

[0044] It is desirable that the vitality engagement model 461, the enthusiasm engagement model 462, and the immersion engagement model 463 are adjusted to receive only the items of explanatory variables that affect their respective outputs among the items of explanatory variables selected from the feature data illustrated in FIG. 1.

[0045] Note that the program installed on each student's smartphone 30 is configured such that each student can specify raw data to be permitted for transmission and raw data not to be permitted, and it is desirable from the viewpoint of personal information protection that only the raw data explicitly permitted is transmitted to the server 20.

[0046] An example of an explanatory variable that can obtain favorable results will be described below. For the vitality engagement model 461, the enthusiasm engagement model 462, and the immersion engagement model 463, by using three feature data items of illuminance data, gyro data, and noise data as explanatory variables, a learning model that can suitably predict student engagement was generated.

[0047] Therefore, by using these three items of feature data as explanatory variables, it is possible to realize an engagement model 46 that outputs a highly accurate target variable while suppressing the amount of calculation. By adding items of explanatory variables, it is possible to realize an engagement model 46 with high prediction accuracy.

[0048] However, the noise data uses the sound detected by the microphone of the smartphone 30. Even if it is described that voices such as conversations are deleted by signal processing, there may be students who feel privacy concerns. It is desirable that students who feel concerned can reject the acquisition of raw data regarding noise.

[0049] Therefore, an engagement model 46 that uses noise data as an explanatory variable and an engagement model 46 that does not use noise data are prepared, and it is preferable to use an engagement model 46 that uses explanatory variables based on raw data permitted by the student himself / herself.

[0050] A combination of illuminance data and inertial data may be used as an explanatory variable. As shown in Table 1, the inertial data may be acceleration data applied to the smartphone 30 measured by an acceleration sensor or angular velocity data applied to the smartphone 30 measured by a gyro sensor. Both acceleration data and angular velocity data may be used simultaneously.

[0051] Charge data regarding whether the battery 38 (see FIG. 3) mounted on the smartphone 30 is being charged or battery remaining amount data may be added as an explanatory variable. Both charge data and battery remaining amount data may be added.

[0052] Position data regarding the position of the smartphone 30, that is, the position of the student, may be added to the explanatory variables. For example, GPS data is used as the raw data for determining the position data.

[0053] When the student has not consented to the transmission of GPS data, the control unit 21 may determine the position data based on the MAC address of the wireless LAN router to which the smartphone 30 is connected. For example, when a router opened by the university for students is used, the control unit 21 determines that the student is at the university. When the same router as the router used almost every night is used, the control unit 21 determines that the student is at home.

[0054] The number of locations where the student stayed during the day may be added to the explanatory variables. The number of locations where the student stayed can be calculated based on GPS data or the MAC address.

[0055] Atmospheric pressure data may be added to the explanatory variables. Screen data regarding whether the screen is lit or not may be added to the explanatory variables. In addition, any combination of feature amount data can be used as the explanatory variables. The combination of feature amount data used as the explanatory variables can be determined, for example, by model compression that simplifies the vitality engagement model 461, the enthusiasm engagement model 462, and the dedication engagement model 463 respectively. Since model compression is well-known, the description thereof is omitted for details.

[0056] The engagement model 46 may be a single learning model that receives an input of feature amount data and outputs comprehensive engagement information. That is, the engagement model 46 may not include the vitality engagement model 461, the enthusiasm engagement model 462, and the dedication engagement model 463.

[0057] FIG. 3 is an explanatory diagram for explaining the configuration of the information processing system 10. The information processing system 10 includes a server 20, a plurality of smartphones 30, and an information processing apparatus 50. The server 20, the smartphones 30, and the information processing apparatus 50 can communicate with each other via a network.

[0058] The server 20 includes a control unit 21, a main memory device 22, an auxiliary storage device 23, a communication unit 24, and a bus. The control unit 21 is an arithmetic control device that executes the program of the present embodiment. One or more CPUs (Central Processing Units), GPUs (Graphics Processing Units), or multi-core CPUs, etc. are used for the control unit 21. The control unit 21 is connected to each hardware component constituting the server 20 via a bus.

[0059] The main memory device 22 is a storage device such as SRAM (Static Random Access Memory), DRAM (Dynamic Random Access Memory), or flash memory. In the main memory device 22, information necessary during the processing performed by the control unit 21 and the program being executed by the control unit 21 are temporarily stored.

[0060] The auxiliary storage device 23 is a storage device such as SRAM, flash memory, hard disk, or magnetic tape. In the auxiliary storage device 23, an estimation result DB (Database) 41, a learning DB 42, an engagement model 46, a program to be executed by the control unit 21, and various data necessary for the execution of the program are stored. The estimation result DB 41 and the learning DB 42 may be stored in an external large-capacity storage device connected to the server 20.

[0061] The communication unit 24 is an interface that performs communication between the server 20 and the network or other devices.

[0062] The server 20 is a mainframe computer, a virtual machine operating on a mainframe computer, a plurality of personal computers performing distributed processing, or a cloud computing system. The server 20 may also be an information device such as a general-purpose personal computer or a tablet.

[0063] The smartphone 30 includes a control unit 31, a main storage device 32, an auxiliary storage device 33, a communication unit 34, a touch panel 35, a sensor 37, a battery 38, and a bus. The touch panel 35 includes an input unit 351 and a display unit 352.

[0064] The control unit 31 is an arithmetic control device that executes the program of the present embodiment. One or more CPUs, GPUs, multi-core CPUs, etc. are used for the control unit 31. The control unit 31 is connected to each hardware part constituting the smartphone 30 via a bus.

[0065] The main storage device 32 is a storage device such as SRAM, DRAM, or flash memory. Information necessary during the processing performed by the control unit 31 and the program being executed by the control unit 31 are temporarily stored in the main storage device 32.

[0066] The auxiliary storage device 33 is a storage device such as SRAM, flash memory, hard disk, or magnetic tape. Programs to be executed by the control unit 31 and various data necessary for the execution of the programs are stored in the auxiliary storage device 33.

[0067] The communication unit 34 is an interface for performing communication between the smartphone 30 and a network or other devices. The display unit 352 is, for example, a liquid crystal display panel or an organic EL (electro-luminescence) panel. The input unit 351 is laminated on the display unit 352. The smartphone 30 may include an external display unit 352 and an input unit 351 such as a mouse.

[0068] The sensor 37 includes, for example, an illuminance sensor used for adjusting the brightness of the display unit 352, a microphone used for voice input, an acceleration sensor, a gyro sensor, a magnetic sensor, a GPS sensor, a barometric pressure sensor, and the like. In the following description, the sensor 37 also includes hardware or software that realizes the function of detecting or acquiring the operating state of the smartphone 30, such as the remaining battery level of the battery 38, the MAC address of the network device to which the communication unit 34 is connected, whether the display unit 352 is lit, and whether the user is operating the smartphone 30.

[0069] Although the illustration of the configuration of the information processing apparatus 50 is omitted, it includes a control unit, a main storage device, an auxiliary storage device, a communication unit, an input unit, a display unit, and the like. The information processing apparatus 50 is a general-purpose information device such as a personal computer, a tablet, or a smartphone. The information processing apparatus 50 may be shared with an information device used by a supporter for other tasks.

[0070] FIG. 4 is an explanatory diagram for explaining the record layout of the estimation result DB 41. The estimation result DB 41 is a database that stores by associating a user ID uniquely assigned to each student, a date, and various engagement information.

[0071] The estimation result DB 41 has a user ID field, a date field, a vitality field, an enthusiasm field, a dedication field, and a comprehensive field. The user ID is recorded in the user ID field. The date is recorded in the date field.

[0072] Vitality engagement information is recorded in the vitality field. Enthusiasm engagement information is recorded in the enthusiasm field. Dedication engagement information is recorded in the dedication field. Comprehensive engagement information is recorded in the comprehensive field. The estimation result DB 41 has one record for one student's one day.

[0073] FIG. 5 is a flowchart for explaining the processing flow of the program. The program in FIG. 5 is executed by the control unit 21 after receiving one-day raw data from the smartphone 30. The control unit 21 creates feature amount data based on the raw data received from the smartphone 30 (step S501).

[0074] The control unit 21 inputs the feature amount data created in step S501 into the engagement model 46 to obtain engagement information (step S502). The control unit 21 creates a new record in the estimation result DB 41 and records the engagement information obtained in step S502 (step S503).

[0075] The control unit 21 determines whether the employment willingness is below a predetermined threshold based on the engagement information obtained in step S502 (step S504). For example, when any one of the vitality engagement information, enthusiasm engagement information, and dedication engagement information is 2 or less, which is the threshold, the control unit 21 determines that the employment willingness is below the threshold. The control unit 21 may also determine whether the employment willingness is below the threshold based on the comprehensive engagement information.

[0076] If it is determined that the employment willingness is below the threshold (YES in step S504), the control unit 21 notifies supporters such as school counselors and the student himself / herself (step S505). If it is determined that the employment willingness is not below the threshold (NO in step S504), or after the end of step S505, the control unit 21 ends the process.

[0077] Note that the state of low employment willingness may be an indication that the student's mental state is not good. Therefore, in step S504, the control unit 21 determines the quality of the student's mental state based on the comparison between the willingness expressed by the engagement information and a predetermined threshold.

[0078] According to this embodiment, an information processing system 10 can be provided that monitors engagement information indicating students' motivation for employment by using the smartphones 30 of individual students. By using the information processing system 10, supporters can intervene early and provide support to students with decreased engagement.

[0079] According to this embodiment, since the smartphone 30 that students usually use is used, an information processing system 10 can be provided that does not impose an additional financial burden on students and their guardians.

[0080] According to this embodiment, after initially installing and setting up the program, individual students do not need to be aware that raw data is being sent to the server 20. Therefore, an information processing system 10 can be realized that monitors employment motivation based on the daily natural state.

[0081] Note that the information processing system 10 of this embodiment can also be used, for example, in educational industries such as vocational schools, cram schools, or preparatory schools, or in organizations such as companies.

[0082] [Embodiment 2] This embodiment relates to an information processing system 10 that uses sleep information as an explanatory variable for the engagement model 46. Regarding the parts common to Embodiment 1, the description will be omitted. Sleep information is information regarding students' sleep every night. Sleep information includes, for example, the bedtime, wake-up time, situation of waking up in the middle of the night, or the quality of sleep.

[0083] As described above, one of the causes of decreased motivation for employment is disruption of lifestyle. Among them, it is known that working night shifts, waking up in the middle of the night, and insomnia are likely to cause a decrease in motivation for employment.

[0084] FIG. 6 is an explanatory diagram for explaining the information processing system 10 of Embodiment 2. Using FIG. 6, the outline of the information processing system 10 that monitors students' motivation for employment will be explained in three stages: the data collection stage, the engagement model generation stage, and the utilization stage.

[0085] [Data collection stage] The data collection stage is almost the same as Embodiment 1 described with reference to FIG. 1. However, in addition to questions regarding employment motivation, the questionnaire includes questions regarding sleep, such as bedtime, wake-up time, and sleep quality.

[0086] [Engagement model generation stage] The control unit 21 generates sleep information including the sleep quality of individual students and the sleep state for each time period based on the questionnaire results. The generation of sleep information may be performed manually by a human. The control unit 21 generates a sleep model 47 by machine learning, using, for example, the feature amount data of the previous night when the student to be monitored answered the questionnaire as the explanatory variable and the generated sleep information as the objective variable.

[0087] The control unit 21 generates an engagement model 46 by machine learning, using the feature amount data for a predetermined period such as the day when the student to be monitored answered the questionnaire or the past three days including the answer date, and the sleep information as the explanatory variables, and the engagement information based on the questionnaire as the objective variable.

[0088] For the machine learning algorithm, a supervised machine learning algorithm for a known classification model such as random forest or CNN is used, for example. Details of the sleep information and the configuration of the engagement model 46 will be described later.

[0089] Note that, regarding the sleep information of the student to be monitored, sleep information measured using a wearable device such as a smartwatch or a bed sensor installed in the bed may be used instead of the questionnaire results. By using objective sleep information as learning data, a highly accurate sleep model 47 can be generated.

[0090] [Utilization stage] Raw data is transmitted from the student's smartphone 30 to the server 20. The control unit 21 converts the raw data transmitted from the smartphone 30 into feature amount data. The control unit 21 inputs the feature amount data into the sleep model 47 to obtain sleep information. The control unit 21 inputs the feature amount data and the sleep information into the engagement model 46 to obtain engagement information.

[0091] The control unit 21 records the sleep information and the engagement information in association with, for example, the ID of each student. Supporters such as school counselors provide support such as recommending that students with sleep disorders such as circadian rhythm reversal or insomnia, who are highly likely to cause a decline in work motivation, visit a medical institution.

[0092] FIG. 7 is an explanatory diagram for explaining the sleep model 47. The sleep model 47 is a learning model generated by machine learning so as to receive an input of an explanatory variable selected from the feature amount data and output sleep information. Here, the sleep information includes the quality of sleep and the sleep state.

[0093] Below FIG. 7, an explanatory diagram for explaining the sleep state is shown. In this figure, the unit time for determining the sleep state is 30 minutes. The downward hatching on the lower right indicates the time zone determined to be in the middle of waking up. The thin downward hatching on the lower left indicates the time zone determined to be during sleep. That is, the sleep information includes classification information that classifies whether the student is in a sleeping state or in the middle of waking up for each time zone separated by the unit time.

[0094] After getting into the futon, the time when sleep first enters is called the sleep onset time. The time when waking state is first entered before getting out of the futon is called the waking time. In the example shown in FIG. 7, there is one middle awakening on Monday, and two middle awakenings on Tuesday, Wednesday, and Thursday.

[0095] Note that the sleep model 47 may be generated separately into a sleep quality model that receives an input of the feature amount data and outputs the quality of sleep, and a sleep state model that receives an input of the feature amount data and outputs the sleep state.

[0096] FIG. 8 is an explanatory diagram for explaining the configuration of the engagement model 46 according to the second embodiment. The engagement model 46 includes a vitality engagement model 461, a passion engagement model 462, and an immersion engagement model 463.

[0097] The vitality engagement model 461 is a learning model that receives explanatory variables selected from feature data and an input of sleep information and outputs vitality engagement information. Similarly, the passion engagement model 462 is a learning model that receives explanatory variables selected from feature data and an input of sleep information and outputs passion engagement information. The immersion engagement model 463 is a learning model that receives explanatory variables selected from feature data and an input of sleep information and outputs immersion engagement information.

[0098] The engagement model 46 outputs engagement information calculated from the vitality engagement information, the passion engagement information, and the immersion engagement information.

[0099] FIG. 9 is a flowchart for explaining the processing flow of the program according to the second embodiment. The program in FIG. 9 is executed by the control unit 21 after receiving one-day raw data from the smartphone 30. The control unit 21 creates feature data based on the raw data received from the smartphone 30 (step S501).

[0100] The control unit 21 inputs the feature data created in step S501 into the sleep model 47 to obtain sleep information (step S512). The control unit 21 inputs the feature data created in step S501 and the sleep information obtained in step S512 into the engagement model 46 to obtain engagement information (step S513).

[0101] The control unit 21 creates a new record in the estimation result DB 41 and records the engagement information acquired in step S513 (step S514). Note that the estimation result DB 41 has a field for recording sleep information, and it is desirable for the control unit 21 to record the sleep information in the estimation result DB 41 together with the engagement information.

[0102] Based on the engagement information acquired in step S502, the control unit 21 determines whether the employment desire is equal to or lower than a predetermined threshold (step S504). If it is determined that the employment desire is equal to or lower than the threshold (YES in step S504), the control unit 21 sends a notification to a supporter such as a school counselor and the student himself / herself (step S505). The supporter confirms the notification using, for example, the information processing device 50. If it is determined that the employment desire is not equal to or lower than the threshold (NO in step S504), or after the completion of step S505, the control unit 21 ends the process.

[0103] According to the present embodiment, by using sleep information, it is possible to provide the information processing system 10 that estimates engagement information with high accuracy.

[0104] Regarding a student who can acquire sleep information measured using a wearable device such as a smartwatch or a bed sensor installed on the bed, instead of using the sleep model 47, the control unit 21 may use the sleep information acquired from the wearable device or the bed sensor to acquire engagement information. By using objective sleep information as an explanatory variable, it is possible to provide the information processing system 10 that estimates engagement information with even higher accuracy.

[0105] [Embodiment 3] This embodiment relates to a method for generating a learning model that generates an engagement model 46 and a sleep model 47. For parts common to Embodiment 2, the description will be omitted.

[0106] FIG. 10 is an explanatory diagram for explaining the record layout of the learning DB 42. The control unit 21 creates the learning DB 42 in the data collection stage described with reference to FIG. 6. The learning DB 42 has a user ID field, a date field, a feature amount data field, and a questionnaire data field.

[0107] The feature amount data field has fields corresponding to various feature amount data exemplified in Table 1, such as an illuminance field, a noise field, an acceleration field, and a gyro field. The data recorded in each field of the feature amount data field will be specifically described. A case where the unit time is 30 minutes will be described as an example. One day is divided into 48 unit times.

[0108] For example, in the field corresponding to the "illuminance" of "2022 / 4 / 10", a sequence of 48 representative values of the illuminance in each unit time arranged in chronological order is recorded. The representative value is a statistical value such as, for example, the average illuminance, the maximum illuminance, the minimum illuminance, the variance, or the standard deviation. A sequence of a plurality of representative values may be recorded respectively.

[0109] Similarly, in the field corresponding to the "noise" of "2022 / 4 / 10", a sequence of 48 representative values of the illuminance in each unit time arranged in chronological order is recorded. The representative value is, for example, the average sound pressure or the maximum sound pressure. The representative value may be a representative value calculated by converting the sound pressure waveform into the frequency domain, such as, for example, the peak frequency.

[0110] The questionnaire data field has a UWES-3 field and a sleep information field. The UWES-3 field has a vitality field, a dedication field, and an absorption field. The sleep information field has a field related to the time when a sleep-related event occurred, such as a bedtime field and a waking field, and a sleep quality field.

[0111] In the UWES-3 field and the sleep information field, the responses of the students to be monitored for each item of the questionnaire are recorded. The learning DB 42 has one record for one day of one student to be monitored.

[0112] Figure 11 is an example screen of the morning questionnaire. The control unit 31 of the smartphone 30 used by the student to be monitored displays the screen shown in Figure 11 every morning. The student to be monitored answers the time to questions regarding the time when events related to sleep occur, such as "What time did you fall asleep last night?" and "What time did you wake up this morning?" The student to be monitored answers a natural number from 1 to 5 to the question regarding the quality of sleep, "Please evaluate the quality of your sleep on a 5-point scale. (1: very bad 5: very good)".

[0113] The control unit 31 transmits the response to the server 20 via the network. The control unit 21 creates a new record in the learning DB 42 and records the received response in the UWES-3 field.

[0114] Figure 12 is an example screen of the evening questionnaire. The control unit 31 of the smartphone 30 used by the student to be monitored displays the screen shown in Figure 12 every night. The student to be monitored selects one of the seven-step buttons from "not at all" to "always feel" for a total of three questions regarding vitality, "I feel energetic when studying (researching).", enthusiasm, "I am enthusiastic about studying (researching).", and immersion, "I am completely immersed in studying (researching).".

[0115] The control unit 31 transmits the response to the server 20 via the network. The control unit 21 records the received response in the sleep information field of the record created in the morning.

[0116] FIG. 13 is a flowchart for explaining the processing flow of a program for generating a learning model. The vitality engagement model 461, the enthusiasm engagement model 462, the dedication engagement model 463, and the engagement model 46 are all generated by a program explained using FIG. 13.

[0117] In the following description, the case where the program shown in FIG. 13 is executed by the control unit 21 will be described as an example. The program shown in FIG. 13 may be executed on hardware separate from the server 20, and each generated model may be stored in the auxiliary storage device 23.

[0118] Using FIG. 13, the processing flow of a program for generating the vitality engagement model 461 using the learning DB 42 will be described. Prior to the execution of the program in FIG. 13, an unlearned model that outputs an objective variable based on explanatory variables is prepared. Further, it is determined which items of the feature data are to be used as explanatory variables.

[0119] The control unit 21 acquires a learning record from the learning DB 42 (step S601). The control unit 21 inputs the explanatory variables included in the acquired learning record into the model being learned and acquires output data (step S602). In the following description, the output data output from the model being learned will be referred to as the output data during learning.

[0120] The control unit 21 adjusts the parameters of the model being learned so that the difference between the data recorded in the vitality field of the learning record acquired in step S601 and the output data during learning becomes small (step S603).

[0121] The control unit 21 determines whether to end the parameter adjustment (step S604). For example, when a predetermined number of learning iterations defined by hyperparameters are repeated, the control unit 21 determines to end the process. The control unit 21 may acquire test data from the learning DB 42, input it into the model being learned, and determine to end the process when an output with a predetermined accuracy is obtained.

[0122] When it is determined not to end the process (NO in step S604), the control unit 21 returns to step S601. When it is determined to end the process (YES in step S604), the control unit 21 records the adjusted parameters in the auxiliary storage device 23 (step S605). Thereafter, the control unit 21 ends the process. Thus, the generation of the vitality engagement model 461 is completed.

[0123] The control unit 21 also generates the enthusiasm engagement model 462 and the dedication engagement model 463 in the same procedure respectively. When generating the enthusiasm engagement model 462, the control unit 21 uses the data recorded in the vitality field as the target variable. When generating the dedication engagement model 463, the control unit 21 uses the data recorded in the dedication field as the target variable.

[0124] The control unit 21 generates the sleep model 47 in the same procedure. When generating the sleep model 47, the control unit 21 generates sleep information that distinguishes whether it is during sleep for each unit time based on data such as the sleep time recorded in the sleep information field. The control unit 21 uses the feature amount data as the explanatory variable and the sleep information and the quality of sleep as the target variables to generate the sleep model 47 by machine learning.

[0125] [Embodiment 4] This embodiment relates to a form of realizing the information processing system 10 by operating in combination a general-purpose computer 90, a program 97, and a smartphone 30. Regarding the parts common to Embodiment 1, the description is omitted.

[0126] FIG. 14 is an explanatory diagram for explaining the configuration of the information processing system 10 according to Embodiment 4. The computer 90 includes a reading unit 29 in addition to the aforementioned control unit 21, main storage device 22, auxiliary storage device 23, communication unit 24, and bus.

[0127] Program 97 is recorded on the portable recording medium 96. The control unit 21 reads program 97 via the reading unit 29 and stores it in the auxiliary storage device 23. Further, the control unit 21 may read program 97 stored in the semiconductor memory 98 such as a flash memory mounted in the computer 90. Furthermore, the control unit 21 may download program 97 from another server computer (not shown) connected via the communication unit 24 and a network (not shown) and store it in the auxiliary storage device 23.

[0128] Program 97 is installed as a control program for the computer 90, loaded into the main memory device 22, and executed. As described above, the server 20 described in the first embodiment is realized. The program 97 of the present embodiment is an example of a program product.

[0129] A computer program can be deployed to be executed on a single computer, or placed at one site, or distributed over multiple sites and executed on multiple computers interconnected by a communication network.

[0130] The technical features (constituent elements) described in each embodiment can be combined with each other, and new technical features can be formed by such combination. The embodiments disclosed this time should be considered as illustrative in all respects and not restrictive. The scope of the present invention is shown not by the above meaning, but by the scope of claims, and is intended to include all modifications within the meaning and scope equivalent to the scope of claims.

[0131] The independent claims and dependent claims recited in the claims can be combined with each other in any combination regardless of the citation format. Furthermore, the claims use, but are not limited to, a format (multi-claim format) that recites claims that cite two or more other claims. A format that recites a multi-claim (multi-multi-claim) that cites at least one multi-claim may also be used.

Description of Signs

[0132] 10 Information processing system 20 Server (computer) 21 Control unit 22 Main memory device 23 Auxiliary storage device 24 Communication unit 29 Reading unit 30 Smartphone 31 Control unit 32 Main memory device 33 Auxiliary storage device 34 Communication unit 35 Touch panel 351 Input unit 352 Display unit 37 Sensor 38 Battery 41 Estimation result DB 42 Learning DB 46 Engagement model 461 Vitality engagement model 462 Enthusiasm engagement model 463 Absorption engagement model 47 Sleep model 50 Information processing device 90 Computer 96 Portable recording medium 97 Program 98 Semiconductor memory

Claims

1. Obtain feature data that can be obtained with the use of a smartphone by a user, Output engagement information regarding engagement when the feature data is input, by inputting the obtained feature data into an engagement model that outputs engagement information regarding engagement when the feature data is input A program for causing a computer to execute the process.

2. Obtain sleep information regarding the user's sleep, The engagement model is a model that outputs engagement information regarding engagement when sleep information and feature data are input, Output engagement information by inputting the obtained sleep information and the obtained feature data into the engagement model The program according to claim 1.

3. The engagement model is A vitality engagement model learned to output vitality engagement information regarding vitality when feature data including illuminance data, inertial data, and charging data and sleep information are input, An enthusiasm engagement model learned to output enthusiasm engagement information regarding enthusiasm when feature data including illuminance data, inertial data, and charging data are input, An immersion engagement model learned to output immersion engagement information regarding immersion when feature data including illuminance data, inertial data, and charging data are input, and includes Output engagement information obtained based on the vitality engagement information, enthusiasm engagement information, and immersion engagement information The program according to claim 2.

4. The sleep information is obtained by inputting the obtained feature data into a sleep model that outputs sleep information regarding sleep when the feature data is input The program according to claim 2.

5. The sleep information includes classification information classifying whether the user is asleep or awake for each time zone The program according to claim 2.

6. The sleep information includes information regarding the quality of sleep The program according to claim 2.

7. The feature data includes illuminance data and inertial data The program according to any one of claims 1 to 6.

8. The feature data includes charging data regarding charging of a battery mounted on the smartphone The program according to claim 7.

9. The feature amount data includes noise data related to noise acquired by the smartphone. The program according to claim 7.

10. The feature amount data includes position data related to the position of the smartphone, atmospheric pressure data acquired by the smartphone, or screen data related to whether or not the screen of the smartphone is lit. The program according to claim 7.

11. The engagement information is a score, Based on a comparison between a predetermined threshold value and the score, the mental state of the user is determined. The program according to claim 1.

12. When the engagement information indicates a decrease in engagement with employment, a notification is output to at least one of a supporter who supports the user and the user. The program according to claim 1.

13. Acquire feature amount data obtained along with the use of the smartphone by the user, By inputting the acquired feature amount data into an engagement model that outputs engagement information related to engagement when the feature amount data is input, engagement information is output. An information processing method executed by a computer.

14. An information processing apparatus including a control unit, The control unit is Acquire feature amount data obtained along with the use of the smartphone by the user, By inputting the acquired feature amount data into an engagement model that outputs engagement information related to engagement when the feature amount data is input, engagement information is output. An information processing apparatus.

15. Acquire learning data from a learning database that records a plurality of sets of feature amount data obtained along with the use of the smartphone by the user and engagement information related to the user in association with each other, Based on the learning data, generate a learning model that outputs engagement information related to engagement when the feature amount data is input. A learning model generation method.

Images