JP7084962B2 - Estimator program, estimator, estimation method, generation program, generation device and generation method - Google Patents

Description

The present invention relates to an estimation program, an estimation device, an estimation method, a generation program, a generation device, and a generation method.

Conventionally, various techniques for dealing with an infectious disease pandemic have been known. For example, a technology that determines whether the sound information includes a cough sound caused by an infectious disease, estimates the number of infected persons from the cough sound caused by an infectious disease, and predicts a pandemic from the number of infected persons, location information, and time information. It has been known.

Japanese Unexamined Patent Publication No. 2019-28687

However, the above-mentioned prior art cannot always estimate the risk of being infected with an infectious disease. For example, in the above-mentioned conventional technique, it is determined whether the sound information includes a cough sound caused by an infectious disease, the number of infected persons is estimated from the cough sound caused by the infectious disease, and the number of infected persons, location information, and time information are obtained. It only predicts the pandemic from. Therefore, it cannot be said that the risk of being infected with an infectious disease can be estimated by the above-mentioned prior art.

The present application is made in view of the above, and proposes an estimation program, an estimation device, an estimation method, a generation program, a generation device, and a generation method capable of estimating the risk of being infected with an infectious disease.

The estimation program according to the present application can estimate the infection risk in the predetermined environment based on the acquisition procedure for acquiring the information on the predetermined environment from the sensor capable of acquiring the information on the predetermined environment and the information on the predetermined environment. It is characterized in that a computer is made to execute an estimation procedure for estimating an infection risk in the surrounding environment of the terminal device by using a machine learning model.

According to one aspect of the embodiment, there is an effect that the risk of being infected with an infectious disease can be estimated.

FIG. 1 is a diagram showing a configuration example of an information processing system according to an embodiment. FIG. 2 is a diagram showing a configuration example of a generator according to an embodiment. FIG. 3 is a diagram showing an example of sensor information according to the embodiment. FIG. 4 is a diagram showing an example of sensor information according to the embodiment. FIG. 5 is a diagram showing an example of sensor information according to the embodiment. FIG. 6 is a diagram for explaining an example of a method for analyzing sensor information according to an embodiment. FIG. 7 is a diagram showing an example of sensor information according to the embodiment. FIG. 8 is a diagram showing an example of sensor information according to the embodiment. FIG. 9 is a diagram showing an example of a learning data set according to an embodiment. FIG. 10 is a diagram for explaining an outline of the generation process according to the embodiment. FIG. 11 is a diagram for explaining an outline of the generation process according to the modified example of the embodiment. FIG. 12 is a diagram showing an example of a learning data set according to a modified example of the embodiment. FIG. 13 is a diagram showing an example of a learning data set according to a modified example of the embodiment. FIG. 14 is a diagram showing an information processing procedure of the generator according to the embodiment. FIG. 15 is a diagram showing a configuration example of the terminal device according to the embodiment. FIG. 16 is a diagram for explaining an example of information processing of the terminal device according to the embodiment. FIG. 17 is a diagram for explaining an example of information processing of the terminal device according to the embodiment. FIG. 18 is a diagram for explaining an example of information processing of the terminal device according to the embodiment. FIG. 19 is a diagram showing an information processing procedure of the terminal device according to the embodiment. FIG. 20 is a hardware configuration diagram showing an example of a computer that realizes the functions of the generation device and the terminal device.

Hereinafter, embodiments (hereinafter, referred to as “embodiments”) for implementing the estimation program, estimation device, estimation method, generation program, generation device, and generation method according to the present application will be described in detail with reference to the drawings. It should be noted that this embodiment does not limit the estimation program, estimation device, estimation method, generation program, generation device and generation method according to the present application. Further, in each of the following embodiments, the same parts are designated by the same reference numerals, and duplicate explanations are omitted.

[1. Introduction]
In recent years, the spread of infectious diseases such as the new coronavirus infection (COVID-19) has become a social problem. In addition, the risk of being infected with an infectious disease (hereinafter, also referred to as infection risk) is a closed space (a closed space with poor ventilation), a dense place (a place where many people are crowded), and a close scene (each other). It is said that it becomes higher when the three conditions (so-called "three dense") of (a scene where conversation and vocalization are performed within reach when reaching out) are overlapped. Therefore, in order to prevent the spread of infectious diseases, it is desirable for each person to avoid an environment with a high risk of infection (so-called "three dense"). However, since the so-called "three dense" is a qualitative index, there is a part that lacks objectivity. Therefore, there is a demand for a technique capable of quantitatively grasping an environment with a high risk of infection (for example, so-called "three dense").

For example, a technique is known in which temperature data is constantly collected through a thermography camera and an alert is notified when hyperthermia is detected (reference URL: https://prtimes.jp/main/html/rd/p/000000003.000054903.html). However, it may not be possible to accurately determine whether or not the temperature data is so-called "three dense". Therefore, for example, it is conceivable to combine various sensor information such as sound data in addition to temperature data to derive an index for determining whether or not it is so-called “three dense”. However, although it is considered that there is some correlation between various sensor information, it is difficult to manually derive the correlation between various sensor information. Therefore, it is expected that it will be difficult to manually derive an index for determining whether or not it is the so-called “three dense”.

Therefore, the generation device 50 according to the embodiment of the present application generates a machine learning model capable of estimating the infection risk in a predetermined environment based on the information about the predetermined environment by machine learning. As a result, the generation device 50 can generate a machine learning model having a high correlation with various sensor information. In addition, the generator 50 can estimate the risk of being infected with an infectious disease.

Further, the terminal device 100 according to the embodiment of the present application uses the machine learning model generated by the generation device 50 to infect the terminal device 100 in the surrounding environment from various sensor information in the peripheral environment of the user. Estimate the risk. Thereby, the terminal device 100 can estimate the risk of being infected with an infectious disease.

(Embodiment)
[2-1. Information processing system configuration example]
First, the configuration of the information processing system 1 according to the embodiment will be described with reference to FIG. FIG. 1 is a diagram showing a configuration example of the information processing system 1 according to the embodiment. The information processing system 1 may include a generation device 50 and a terminal device 100. The generation device 50 may be communicably connected to the terminal device 100 via a predetermined network N. The generation device 50 may be communicably connected to the terminal device 100 by wire or wirelessly. The information processing system 1 may include an arbitrary number of generation devices 50 and an arbitrary number of terminal devices 100.

Hereinafter, the user specified by the user ID "U1" may be described as "user U1". In the following, when "user U * (* is an arbitrary numerical value)" is described, it may indicate that the user is a user specified by the user ID "U *". For example, when the description is "user U2", the user may be a user specified by the user ID "U2".

Further, in the following, the terminal device 100 may be described as the terminal devices 100 ₁ and 100 ₂ depending on the user who uses the terminal device 100. For example, the terminal device 100 ₁ may be the terminal device 100 used by the user U1. Further, for example, the terminal device 100 ₂ may be the terminal device 100 used by the user U2. Further, in the following, when the terminal devices 100 ₁ and 100 ₂ are described without particular distinction, they may be described as the terminal device 100.

The generation device 50 can generate a machine learning model capable of estimating the infection risk in a predetermined environment based on information about a predetermined environment (for example, various sensor information acquired in the predetermined environment by various sensors). It may be an information processing device. First, the generation device 50 may acquire, for example, various sensor information acquired by the terminal device 100 from the terminal device 100. For example, the generation device 50 may acquire sensor information regarding the surrounding environment of the terminal device 100. Subsequently, the generation device 50 may generate a machine learning model capable of estimating the infection risk in the surrounding environment of the terminal device 100 based on the acquired sensor information. Further, the generation device 50 may distribute information about the generated machine learning model to the terminal device 100.

The terminal device 100 may be an information processing device used by the user. The terminal device 100 may be realized by, for example, a smartphone, a tablet terminal, a notebook PC (Personal Computer), a desktop PC, a mobile phone, a PDA (Personal Digital Assistant), or the like.

Further, the terminal device 100 can acquire information about the machine learning model generated by the generation device 50 from the generation device 50. For example, the terminal device 100 can acquire a machine learning model capable of estimating the infection risk in a predetermined environment based on the information about the predetermined environment.

Further, the terminal device 100 may include various sensors. For example, the terminal device 100 may include a temperature sensor (also referred to as a motion sensor), an image sensor (camera), a sound sensor, and a wind speed sensor. Further, the terminal device 100 can acquire various sensor information in a predetermined environment by various sensors. For example, the terminal device 100 can acquire various sensor information regarding the surrounding environment of the terminal device 100 by various sensors. The various sensors may be configured by a conventionally known method.

In addition, the terminal device 100 can estimate the risk of infection in the surrounding environment of the terminal device 100. For example, the terminal device 100 can estimate the infection risk in the surrounding environment of the terminal device 100 based on the sensor information regarding the surrounding environment of the terminal device 100 by using the machine learning model acquired from the generation device 50. Thereby, the terminal device 100 can estimate the infection risk at the user's current position.

Further, the terminal device 100 can output the estimation result of the infection risk in the surrounding environment of the terminal device 100. As a result, the terminal device 100 can notify the user of the estimation result of the infection risk at the user's current position. Further, the terminal device 100 may output the estimation result of the terminal device of another user. As a result, the user of the terminal device 100 can grasp the risk of infection in the surrounding environment of other users.

[2-2. Configuration example of generator]
Next, the configuration of the generation device 50 according to the embodiment will be described with reference to FIG. FIG. 2 is a diagram showing a configuration example of the generation device 50 according to the embodiment. The generation device 50 may have a communication unit 51, a storage unit 52, and a control unit 53. The generation device 50 may have an input unit (for example, a keyboard, a mouse, etc.) that receives various operations from the administrator of the generation device 50 or the like.

(Communication unit 51)
The communication unit 51 may be realized by, for example, a NIC (Network Interface Card) or the like. The communication unit 51 may transmit / receive information to / from the terminal device 100, for example. The communication unit 51 may be connected to the network N by wire or wirelessly. The communication unit 51 may be able to communicate with any external device by wire or wirelessly.

(Memory unit 52)
The storage unit 52 may be realized by, for example, a semiconductor memory element such as a RAM (Random Access Memory) or a flash memory (Flash Memory), or a storage device such as a hard disk or an optical disk. The storage unit 52 can store various programs (corresponding to an example of an information processing program). For example, the storage unit 52 can store the generation program according to the embodiment. Further, the storage unit 52 can store various data. For example, the storage unit 52 can store the sensor information acquired by the acquisition unit 531. Further, the storage unit 52 can store the learning data of the machine learning model generated by the generation unit 532. Further, the storage unit 52 can store the model data of the trained machine learning model generated by the generation unit 532.

Here, the generation program according to the embodiment may be a program that realizes the function of the control unit 53. For example, the generation program according to the embodiment may cause a computer to execute an acquisition procedure for acquiring information on a predetermined environment. Further, for example, the generation program according to the embodiment may cause a computer to execute a generation procedure for generating a machine learning model capable of estimating the infection risk in the surrounding environment of the terminal device based on information on a predetermined environment.

(Control unit 53)
The control unit 53 may be a controller of the generation device 50. The control unit 53 may be configured by, for example, a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or the like. The control unit 53 may execute various programs (corresponding to an example of an information processing program) stored in the storage device inside the generation device 50 using the RAM as a work area. For example, the control unit 53 may execute the generation program according to the embodiment by using the RAM as a work area by the CPU, MPU, or the like. Further, the control unit 53 may be realized by an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

The control unit 53 may include an acquisition unit 531, a generation unit 532, and a distribution unit 533. In one embodiment, the control unit 53 can realize or execute the information processing operation described below. The internal configuration of the control unit 53 is not limited to the configuration shown in FIG. 2, and may be any other configuration as long as it is configured to perform information processing described later.

(Acquisition unit 531)
The acquisition unit 531 can acquire information about a predetermined environment. Specifically, the acquisition unit 531 can acquire the sensor information acquired by the sensor capable of acquiring the sensor information regarding the predetermined environment as an example of the information regarding the predetermined environment. For example, the acquisition unit 531 can acquire the sensor information acquired by the terminal device 100 having a sensor capable of acquiring the sensor information regarding a predetermined environment. For example, the acquisition unit 531 can acquire sensor information regarding the surrounding environment of the user's terminal device 100 as an example of information regarding the surrounding environment of the user's terminal device 100. For example, the acquisition unit 531 may acquire sensor information from a terminal device 100 having a sensor capable of acquiring sensor information regarding a predetermined environment. Further, for example, the acquisition unit 531 acquires sensor information from a device other than the terminal device 100 such as a database in which the sensor information acquired by the terminal device 100 having a sensor capable of acquiring sensor information regarding a predetermined environment is stored. You may. More specifically, the acquisition unit 531 uses, for example, the sensor information acquired by the terminal device 100 together with the position information indicating the position where the sensor information is acquired and the time information indicating the date and time when the sensor information is acquired. It may be obtained from 100. The acquisition unit 531 may store the acquired sensor information in the storage unit 52 in association with the position information and the time information.

For example, the acquisition unit 531 can acquire the temperature information in the surrounding environment of the terminal device 100 acquired by the temperature sensor as an example of the sensor information. This point will be described with reference to FIG. FIG. 3 is a diagram showing an example of sensor information according to the embodiment. As an example of the sensor information, the acquisition unit 531 can acquire a temperature distribution image in the surrounding environment of 100 terminals mounted by a thermography camera. In the temperature distribution image G11 exemplified in FIG. 3, the low temperature region is shown by R11, the medium temperature region is shown by R12, and the high temperature region is shown by R13. Specifically, in the temperature distribution image G11, for example, when the human body is regarded as a high temperature object, the region where people are densely shown is shown as a high temperature region in the temperature distribution image G11. The acquisition unit 531 may acquire information on the degree of crowding in the surrounding environment of the terminal device 100 from the distribution pattern in the high temperature region in the temperature distribution image G11.

The acquisition unit 531 may acquire only the temperature information assumed as the body temperature of the human body among the temperature information acquired by the temperature sensor. For example, the acquisition unit 531 may remove temperature information other than the temperature information assumed as the body temperature of the human body from the temperature information acquired by the temperature sensor.

For example, the generation unit 532 may train the machine learning model to estimate the infection risk based on the distribution pattern in the high temperature region of the temperature distribution image G11 acquired by the acquisition unit 531. For example, when it is estimated from the distribution pattern of the high temperature region of the temperature distribution image G11 that the crowd is low (for example, when the distribution pattern of the high temperature region is sparse), the generation unit 532 sets the risk of infection. You may train the machine learning model to underestimate. Further, the generation unit 532 is infected when it is estimated from the distribution pattern of the high temperature region of the temperature distribution image that the crowd is crowded (for example, when the distribution pattern of the high temperature region is locally concentrated). You may train the machine learning model to estimate the risk high.

Further, as an example of the sensor information, the acquisition unit 531 can acquire the temperature information in the surrounding environment of the terminal device 100 acquired by the temperature sensor at predetermined time intervals. This point will be described with reference to FIG. FIG. 4 is a diagram showing an example of sensor information according to the embodiment. In the example shown in FIG. 4, the acquisition unit 531 can acquire a temperature distribution image in the surrounding environment of the terminal device 100 taken by a thermography camera at predetermined time intervals as an example of the sensor information. The acquisition unit 531 may acquire information regarding the presence or absence of human retention in the surrounding environment of the terminal device 100 from the distribution pattern of each high temperature region of the temperature distribution image in the surrounding environment of the terminal device 100 taken at predetermined time intervals. ..

For example, the temperature distribution image G22 shown in the upper part of FIG. 4 is an image taken T seconds after the temperature distribution image G21 is taken. Further, the temperature distribution image G23 is an image taken T seconds after the temperature distribution image G22 is taken. In the example shown in the upper part of FIG. 4, no significant change is observed in the distribution in the high temperature region during the temporal transition from the temperature distribution image G21 to G22 and from G22 to G23. Therefore, the acquisition unit 531 may acquire information indicating that there is a human retention, based on, for example, the temperature distribution images G21 to G23.

For example, the temperature distribution images G21 to G23 will be described by dividing them into a total of 16 sections of 4 vertical × 4 horizontal. In this case, the high temperature region of the temperature distribution image G21 is 7 out of 16 compartments, whereas the high temperature region of the temperature distribution image G22 is 6 out of 16 compartments. That is, in this case, the high temperature region of the temperature distribution image G22 is reduced by one section as compared with the temperature distribution image G21. Therefore, for example, according to the temperature distribution image G22, it is shown that almost no movement of people occurs during T seconds from the crowded state shown by the temperature distribution image G21. That is, according to the temperature distribution image G22, for example, it is shown that the crowded state of the same degree as the crowd shown by the temperature distribution image G21 was maintained after T seconds.

Further, the high temperature region of the temperature distribution image G22 is 6 out of 16 compartments, while the high temperature region of the temperature distribution image G23 is 7 out of 16 compartments. That is, the high temperature region of the temperature distribution image G23 is increased by only one section as compared with the temperature distribution image G22. For example, according to the temperature distribution image G23, it is shown that almost no movement of people occurred during T seconds from the crowded state shown by the temperature distribution image G22. That is, for example, according to the temperature distribution image G23, it is shown that the crowded state of the same degree as the crowd shown by the temperature distribution image G22 was maintained after T seconds. Therefore, in this case, the acquisition unit 531 may acquire information indicating that there is a human retention based on the temperature distribution images G21 to G23.

Further, the temperature distribution image G25 shown in the lower part of FIG. 4 is an image taken T seconds after the temperature distribution image G24 was taken. Further, the temperature distribution image G26 is an image taken T seconds after the temperature distribution image G25 is taken. In the example shown in the lower part of FIG. 4, a large change is observed in the distribution in the high temperature region when the temperature distribution image G24 to G25 and G25 to G26 are temporally transitioned. Therefore, the acquisition unit 531 may acquire information indicating that there is no human retention, based on, for example, the temperature distribution images G24 to G26.

For example, the temperature distribution images G24 to G26 will be described by dividing them into a total of 16 sections of 4 vertical × 4 horizontal. In this case, the high temperature region of the temperature distribution image G24 is 7 out of 16 compartments, whereas the high temperature region of the temperature distribution image G25 is 2 out of 16 compartments. That is, the high temperature region of the temperature distribution image G25 is reduced by 5 sections as compared with the temperature distribution image G24. Therefore, for example, according to the temperature distribution image G25, it is shown that the movement of people occurred during T seconds from the crowded state shown by the temperature distribution image G24. That is, according to the temperature distribution image G25, for example, it is shown that after T seconds, the number of people is smaller than that shown in the temperature distribution image G24.

Further, the high temperature region of the temperature distribution image G25 is 2 out of 16 compartments, while the high temperature region of the temperature distribution image G26 is 7 out of 16 compartments. That is, the high temperature region of the temperature distribution image G26 is increased by 5 sections as compared with the temperature distribution image G25. Therefore, for example, according to the temperature distribution image G26, it is shown that the movement of people has occurred again within T seconds from the crowded state shown by the temperature distribution image G25. That is, according to the temperature distribution image G26, for example, it is shown that after T seconds, the crowded state is larger than that shown by the temperature distribution image G25. Therefore, in this case, the acquisition unit 531 may acquire information indicating that there is no human retention based on the temperature distribution images G24 to G26.

The generation unit 532 may train the machine learning model to estimate the infection risk based on the distribution pattern in the high temperature region of the temperature distribution image, for example. Further, the generation unit 532 may train the machine learning model to estimate the infection risk based on the distribution pattern of each high temperature region of the temperature distribution image in the surrounding environment of the terminal device 100 taken at a predetermined time interval, for example. .. For example, the generation unit 532 may train a machine learning model to estimate the risk of infection high based on the information indicating that there is a human retention. In addition, the generation unit 532 may train the machine learning model so as to estimate the infection risk low based on the information indicating that there is no human retention.

Further, as an example of the sensor information, the acquisition unit 531 may acquire the movement of a human figure included in the image in the surrounding environment of the terminal device 100 taken by the camera within a predetermined time. Generally, you want your child to play in a park with good weather, but you may be concerned about the risk of infection. However, acquiring a face image of a person from another place with a camera in order to acquire information on crowds is not possible from the viewpoint of privacy. Therefore, the acquisition unit 531 can acquire only an image of a human figure without acquiring a human face image or the like. This point will be described with reference to FIG.

FIG. 5 is a diagram showing an example of sensor information according to the embodiment. For example, the acquisition unit 531 can acquire the image G31 of the human figure by extracting only the portion of the human figure from the image in the surrounding environment of the terminal device 100 taken by the camera. The black region S1 of the image G31 shows that human figures are concentrated in one place. Further, the acquisition unit 531 can acquire the image G32 of another human figure by extracting only the portion of the human figure from another image taken by the camera within a predetermined time. The region S2 shown in black in the image G32 shows one of a plurality of dispersely dispersed figures. In this way, the acquisition unit 531 can acquire information on the degree of crowding of people in the surrounding environment of the terminal device 100 by tracking the movement of a human figure included in the image captured within a predetermined time. For example, the acquisition unit 531 may set a predetermined number of pixels as a threshold value. In this case, the acquisition unit 531 may acquire information indicating that people are crowded when, for example, the number of pixels of the image G32 of the human figure satisfies a predetermined threshold value. Further, the acquisition unit 531 may acquire information indicating that people are not crowded when, for example, the number of pixels of the image G32 of the human figure does not satisfy a predetermined threshold value.

The acquisition unit 531 may acquire only the image of a human figure among the images taken by the camera. For example, the acquisition unit 531 may remove a shadow other than a shadow assumed as a human shadow (for example, a shadow of a tree in a park, a shadow of a building, etc.) from an image taken by a camera.

At this time, for example, the generation unit 532 may train the machine learning model to estimate the infection risk based on the movement of the human figure included in the image in the surrounding environment of the terminal device 100 taken within a predetermined time. For example, when the generation unit 532 is crowded with many people (for example, when the pattern changes from a pattern in which a plurality of shadows are dispersed apart to a pattern in which the shadows are concentrated in one place after a lapse of a predetermined time). May train a machine learning model to estimate the risk of infection high. That is, the generation unit 532 may train the machine learning model so as to estimate the infection risk high based on the information indicating that the people are crowded, for example. In addition, when the generation unit 532 is not crowded with people (for example, from a pattern in which a plurality of shadows are scattered separately, a pattern in which a plurality of shadows are scattered even after a lapse of a predetermined time remains). May be trained in a machine learning model to underestimate the risk of infection. That is, the generation unit 532 may train a machine learning model to underestimate the risk of infection, for example, based on information indicating that people are not crowded.

Here, an example of a method for analyzing sensor information according to an embodiment will be described with reference to FIG. FIG. 6 is a diagram showing an example of a method for analyzing sensor information according to an embodiment. In the example shown in FIG. 6, an example of the analysis method of the image G3 of the human figure carried out by the control unit 53 will be described. First, the control unit 53 may subdivide the image G3 of a human figure in one frame into a grid pattern, for example. Next, the control unit 53 may digitize the color information for each grid. Then, the control unit 53 may create a numerical matrix in which the numerical values of the color information are associated with each grid. For example, the control unit 53 may create a numerical matrix in which the numerical value "0" is associated with the white grid and the numerical value "1" is associated with the black grid corresponding to the region S3 shown in black. The control unit 53 may, for example, create a numerical matrix in which numerical values are arranged in the direction of the arrow in FIG. Further, when the control unit 53 analyzes the movement of the shadow within the predetermined time, the control unit 53 may create a numerical matrix for each image for the number of frames captured within the predetermined time.

Further, the analysis method shown in FIG. 6 may be applied to a temperature distribution image. For example, the control unit 53 may subdivide the temperature distribution image G11 of one frame into a grid pattern. The control unit 53 may quantify the temperature information of each grid. The control unit 53 may create a numerical matrix in which the numerical values of the temperature information are associated with each grid. For example, the control unit 53 creates a numerical matrix in which the numerical value "0" is associated with the grid in the low temperature region, the numerical value "1" is associated with the grid in the medium temperature region, and the numerical value "2" is associated with the grid in the high temperature region. It's okay. Further, the control unit 53 may create a numerical matrix for each image of the number of frames taken at the predetermined time interval for the temperature distribution image of the crowd taken at the predetermined time interval shown in FIG.

The generation unit 532 may generate teacher data used in supervised learning. For example, the generation unit 532 may generate data labeled according to the high risk of infection for each numerical matrix as teacher data. For example, the generation unit 532 may attach the label "danger" or the numerical label "3" to the numerical matrix relating to the environment where the risk of infection is relatively high. In addition, the generation unit 532 may attach the label "Caution" or the numerical label "2" to the numerical matrix relating to the environment where the risk of infection is medium. In addition, the generation unit 532 may attach the label "OK" or the numerical label "1" to the numerical matrix relating to the environment where the risk of infection is relatively low.

Further, the acquisition unit 531 can acquire the sound information in the surrounding environment of the terminal device 100 acquired by the sound sensor as an example of the sensor information. This point will be described with reference to FIG. 7. As an example of sound information, the acquisition unit 531 may acquire information indicating a time change in the volume of a human voice in the surrounding environment of the terminal device 100. In FIG. 7, for the sake of explanation, the information on the time change of the volume is illustrated as a graph (hereinafter, also referred to as a volume graph) G4. The volume graph G4 may be divided into three areas: a volume graph G41 in the range of time t0 to t1, a volume graph G42 in the range of time t1 to t2, and a volume graph G43 in the range after time t2. The volume graph G41 is a region where the volume is low and the volume hardly changes. The volume graph G42 is an area where the volume changes from a low volume state to a high volume state. The volume graph G43 is a region where the volume is large and the change in volume is large.

The generation unit 532 may train the machine learning model to estimate the infection risk based on the pattern of the time change of the volume in the surrounding environment of the terminal device 100 acquired within a predetermined time, for example. For example, the generation unit 532 trains a machine learning model to estimate the risk of infection high when a person makes a loud voice nearby (for example, in the case of a pattern similar to the region of the volume graph G43). good. In addition, the generator 532 is machine-learned to moderately estimate the risk of infection when it is estimated that a person is approaching while speaking (for example, in the case of a pattern similar to the region of the volume graph G42). You may train the model. Further, the generation unit 532 trains a machine learning model to estimate the infection risk low when a person makes a small voice in a distant place (for example, in the case of a pattern similar to the region of the volume graph G41). good.

The acquisition unit 531 may acquire only the human voice information among the sound information acquired by the sound sensor. For example, the acquisition unit 531 is not caused by the movement of a person from the sound information acquired by the sound sensor, such as a sound other than the sound assumed as a human voice (for example, a machine driving sound, a leaf rubbing sound, etc.). Noise) may be removed. The acquisition unit 531 may acquire information on sounds caused by movements of people other than human voices. For example, the acquisition unit 531 may acquire information on sounds such as shoe sounds and clothes rubbing in the surrounding environment of the terminal device 100. Further, the acquisition unit 531 may acquire the frequency information of the sound in the surrounding environment of the terminal device 100 as an example of the sound information. The generation unit 532 may determine whether or not the sound is caused by the movement of a person, for example, by analyzing the frequency of the sound information. The generation unit 532 may perform frequency analysis by any conventionally known method such as FFT (Fast Fourier Transform). In this case, the generation unit 532 may train the machine learning model to estimate the infection risk based on the frequency information of the sound caused by the movement of the person.

As an example of the sensor information, the acquisition unit 531 can acquire the wind speed information in the surrounding environment of the terminal device 100 acquired by the wind speed sensor. This point will be described with reference to FIG. The acquisition unit 531 can acquire information indicating the time change of the wind speed. In FIG. 8, for the sake of explanation, information indicating the time change of the wind speed is shown as a graph (hereinafter, also referred to as a wind speed graph) G5. The wind speed graph G5 may be divided into three regions: a wind speed graph G51 in the range of time t0 to t3, a wind speed graph G52 in the range of time t3 to t4, and a wind speed graph G53 in the range after time t4. The wind speed graph G51 is a region where the wind speed is large and the change in the wind speed is large. The wind speed graph G52 is a region where the wind speed changes from a high wind speed state to a low wind speed state. The wind speed graph G53 is a region where the wind speed is small and there is almost no change in the wind speed.

The generation unit 532 may train the machine learning model to estimate the infection risk based on the pattern of the time change of the wind speed in the surrounding environment of the terminal device 100 acquired within a predetermined time, for example. For example, the generator 532 has a low risk of infection when the environment has good air flow (good air circulation and sufficient ventilation) (for example, in the case of a pattern similar to the region of the wind speed graph G51). You may train the machine learning model to make an estimate. In addition, the generator 532 has a moderate risk of infection in an environment where the air flow is getting worse (ventilation is getting worse) (for example, in the case of a pattern similar to the region of the wind speed graph G52). You may train the machine learning model to make an estimation. Further, the generation unit 532 has a high risk of infection in an environment where the air flow is poor (air is stagnant and ventilation is poor) (for example, in the case of a pattern similar to the region of the wind speed graph G53). You may train the machine learning model to make an estimate.

(Generation unit 532)
The generation unit 532 can generate a machine learning model capable of estimating the infection risk in the surrounding environment of the terminal device 100 based on the information about the predetermined environment. For example, the generation unit 532 can generate a machine learning model that estimates the infection risk in the surrounding environment of the terminal device 100 based on the sensor information acquired by the acquisition unit 531. For example, the generation unit 532 can train a machine learning model by clustering, which is a kind of unsupervised learning. This point will be specifically described with reference to FIGS. 9 and 10.

First, the learning data set according to the embodiment will be described with reference to FIG. FIG. 9 is a diagram showing an example of a learning data set according to an embodiment. The learning data set 221A may have items such as "data set ID", "crowd", "retention", "shadow", "sound", and "wind".

In the learning data set 221A, the item of "data set ID" may store, for example, identification information for identifying the learning data set consisting of a set of various sensor information acquired by the terminal device 100. Here, the learning data set identified by the data set ID may be a set of sensor information regarding the same surrounding environment of the terminal device 100 of the same user. Further, the learning data set identified by the data set ID may be a set of sensor information acquired at the same position at the same date and time. These may be the same in the modification of the learning data set described later.

In the training data set 221A, the item of "crowd" may store information indicating a temperature distribution image of the crowd. For example, in the item of "crowd", the RGB numerical value for each pixel of the temperature distribution image of one frame may be stored.

In the training data set 221A, the item of "retention" may store information indicating a set of temperature distribution images of crowds taken at predetermined time intervals. For example, in the item of "retention", RGB numerical values for each pixel of each temperature distribution image may be stored in association with the image for each temperature distribution image for the number of frames taken at a predetermined time interval.

In the training data set 221A, the item of "shadow" may store information indicating a set of images of a human figure taken within a predetermined time. For example, in the "shadow" item, for each image of a human figure taken within a predetermined time for the number of frames, a numerical value corresponding to black and white for each pixel of each image of the human image (white is "0", black is "0"). "1") may be stored in association with the image.

In the learning data set 221A, the volume information acquired within a predetermined time may be stored in the item of "sound". For example, in the item of "sound", measured values such as the instantaneous maximum volume and the average volume within the predetermined time may be stored as the volume information acquired within the predetermined time.

In the learning data set 221A, the item of "wind" may store the wind speed information acquired within a predetermined time. For example, in the item of "wind", measured values such as an instantaneous maximum wind speed, an average wind speed, and a wind direction within a predetermined time may be stored as wind speed information acquired within a predetermined time.

Next, the generation process according to the embodiment will be described with reference to FIG. FIG. 10 is a diagram for explaining an outline of the generation process according to the embodiment. When sensor information is input to the machine learning model, the generation unit 532 performs machine learning so that the input sensor information is classified into at least one of the three clusters according to the high risk of infection. You can train the model. For example, when the generation unit 532 uses the k-means method and sensor information is input to the machine learning model, the input sensor information is at least one of three clusters according to the high risk of infection. Machine learning models can be trained to be classified into one cluster. The generation unit 532 is not limited to the k-means method, and may train a machine learning model by using various clustering methods such as logistic regression using a Dirichlet process.

For example, the generation unit 532 can train a machine learning model based on the training data set 221A. For example, the generation unit 532 is a machine learning model based on m learning data sets (for example, m is a natural number of several thousand to tens of thousands) identified by the data sets IDs “DS-1C” to “DS-mC”. Can be learned. Here, it is assumed that each training data set is N-dimensional data. The N dimension may mean that the number of data entries (numerical values) in one data set is N. Further, when the data for a certain item, for example, "shadow" cannot be acquired, the generation unit 532 does not have to use the data set for learning. Further, when the data for a certain item cannot be acquired, the generation unit 532 may set all the numerical values of the data entries of the item to 0 and use it for learning.

For example, when m learning data sets identified by the data set IDs "DS-1C" to "DS-mC" are input to the machine learning model, the generation unit 532 is a sensor related to an environment having a relatively high risk of infection. Machine learning to be classified into at least one of the cluster C1 to which the information belongs, the cluster C2 to which the sensor information about the environment with a medium infection risk belongs, and the cluster C3 to which the sensor information about the environment with a relatively low infection risk belongs. You can train the model. As described above, when the sensor information is input to the machine learning model, the generation unit 532 classifies the sensor information into at least one of two or more clusters according to the high risk of infection. Generate a trained machine learning model. Further, the generation unit 532 may train the machine learning model so that the input sensor information outputs a label corresponding to the classified cluster as the output information of the machine learning model.

For example, the generation unit 532 may set the label corresponding to the cluster C1 to "danger", the label corresponding to the cluster C2 to be "caution", and the label corresponding to the cluster C3 to be "OK". At this time, when the training data set input to the machine learning model is classified into the cluster C1, the generation unit 532 outputs the label "danger" corresponding to the cluster C1 as the output information of the machine learning model. The learning model may be trained. Further, when the training data set input to the machine learning model is classified into the cluster C2, the generation unit 532 is machine learning so as to output the label "Caution" corresponding to the cluster C2 as the output information of the machine learning model. You may train the model. Further, when the training data set input to the machine learning model is classified into the cluster C3, the generation unit 532 is machine learning so as to output the label "OK" corresponding to the cluster C3 as the output information of the machine learning model. You may train the model. As described above, when the sensor information is input as the input information of the machine learning model, the generation unit 532 outputs a label indicating the infection risk corresponding to the cluster in which the sensor information is classified as the output information of the machine learning model. Generate a machine learning model trained to do.

The generation unit 532 may compress the clusters in the N-dimensional space to a lower dimension by using principal component analysis (PCA). This point will be specifically described with reference to FIG. FIG. 11 is a diagram for explaining an outline of the generation process according to the modified example of the embodiment. The generation unit 532 can make the calculation relatively easy by compressing the dimensions of the cluster in the N-dimensional space by using the principal component analysis (PCA). For example, as shown on the upper right side of FIG. 11, the generation unit 532 uses principal component analysis to generate two-dimensional data (black circles) composed of a feature amount “crowd” and another feature amount “shadow”. (Points indicated by) can be compressed into one-dimensional data (points indicated by white circles). As described above, the generation device 50 may perform the calculation by compressing the N-dimensional feature amount to an arbitrary dimension by using the principal component analysis. The generation device 50 may, for example, compress an N-dimensional feature amount to two dimensions for calculation.

In the above embodiment, an example in which the generation unit 532 learns a machine learning model by clustering, which is a kind of unsupervised learning, has been described, but the machine learning method is not limited to this. For example, the generation unit 532 may train a machine learning model by supervised learning. For example, the generation unit 532 may train a machine learning model based on a learning data set which is a set of sensor information and a teacher label. For example, the generation unit 532 may train the machine learning model so that the machine learning model outputs a teacher label corresponding to the sensor information when the sensor information is input to the machine learning model. This point will be specifically described with reference to FIGS. 12 and 13.

First, the learning data set according to the modified example of the embodiment will be described with reference to FIG. FIG. 12 is a diagram showing an example of a learning data set according to a modified example of the embodiment. The learning data set 221B may have items such as "data set ID", "crowd", "retention", "shadow", "sound", "wind", and "teacher label".

In the learning data set 221B, the item of "data set ID" may store, for example, identification information for identifying the learning data set consisting of a set of various sensor information acquired by the terminal device 100 and a teacher label.

In the training data set 221B, the item of "crowd" may store, for example, information indicating a temperature distribution image of the crowd. For example, in the item of "crowd", the RGB numerical value for each pixel of the temperature distribution image of one frame may be stored.

In the training data set 221B, the item of "retention" may store, for example, information indicating a set of temperature distribution images of crowds taken at predetermined time intervals. For example, in the item of "retention", RGB numerical values for each pixel of each temperature distribution image may be stored in association with the image for each temperature distribution image for the number of frames taken at a predetermined time interval.

In the training data set 221B, the item of "shadow" may store, for example, information indicating a set of images of a human figure taken within a predetermined time. For example, in the "shadow" item, for each image of a human figure taken within a predetermined time for the number of frames, a numerical value corresponding to black and white for each pixel of each image of the human image (white is "0", black is "0"). "1") may be stored in association with the image.

In the training data set 221B, the item of "sound" indicates, for example, the volume information acquired within a predetermined time. For example, in the item of "sound", measured values such as the instantaneous maximum volume and the average volume within the predetermined time may be stored as the volume information acquired within the predetermined time.

In the training data set 221B, the item of "wind" indicates, for example, wind speed information acquired within a predetermined time. For example, in the item of "wind", measured values such as an instantaneous maximum wind speed, an average wind speed, and a wind direction within a predetermined time may be stored as wind speed information acquired within a predetermined time.

In the training data set 221B, the item of "teacher label" may store, for example, information indicating the high risk of infection in the environment regarding the set of sensor information identified by the data set ID. For example, the teacher label "3" indicates that it is sensor information about an environment where the risk of infection is relatively high. The teacher label "2" also indicates that the sensor information is for an environment with a moderate risk of infection. Further, the teacher label "1" indicates that the sensor information is related to the environment where the risk of infection is relatively low.

For example, the generation unit 532 can train a machine learning model based on the training data set 221B. For example, the generation unit 532 is a machine learning model based on m learning data sets (for example, m is a natural number of several thousand to tens of thousands) identified by the data sets IDs “DS-1A” to “DS-mA”. Can be learned. For example, the generation unit 532 inputs a set of sensor information (crowd # 1, retention # 1, shadow # 1, sound # 1, wind # 1) identified by the data set ID "DS-1A" into the machine learning model. If so, the machine learning model may be trained to output the teacher label "3". Further, the generation unit 532 inputs a set of sensor information (crowd # 2, retention # 2, shadow # 2, sound # 2, wind # 2) identified by the data set ID “DS-2A” into the machine learning model. If so, the machine learning model may be trained to output the teacher label "2". In this way, the generation unit 532 generates a trained machine learning model by training m training data sets identified by the data sets IDs "DS-1A" to "DS-mA". good.

Next, the learning data set according to the modified example of the embodiment will be described with reference to FIG. FIG. 13 is a diagram showing an example of a learning data set according to a modified example of the embodiment. In the learning data set according to the modified example, each sensor information may be given a score according to the high infection risk estimated from each sensor information. Specifically, the score is "3" for the sensor information related to the environment with a relatively high infection risk, the score is "2" for the sensor information related to the environment with a medium infection risk, and the sensor information related to the environment with a relatively low infection risk. May be given a score of "1". Then, the teacher label may be given based on the total of the points given to the various sensor information. For example, when the total score of various sensor information exceeds 12, a teacher label "3" indicating that the sensor information is related to an environment having a relatively high infection risk may be given. Further, when the total score of the various sensor information is 9 to 12 points, the teacher label "2" indicating that the sensor information is related to the environment where the infection risk is moderate may be given. Further, when the total score of various sensor information is 9 points or less, a teacher label "1" indicating that the sensor information is related to an environment having a relatively low infection risk may be given. Of the various sensor information, the specific sensor information may be weighted to calculate the total score of the various sensor information.

The learning data set 221C according to the modification may have items such as "data set ID", "crowd", "retention", "shadow", "sound", "wind", and "teacher label".

In the training data set 221C, in the item of "data set ID", for example, identification information for identifying the training data set consisting of a set of points and a teacher label given to various sensor information acquired by the terminal device 100 is included. May be stored.

In the training data set 221C, the item of "crowd" may store, for example, information indicating the high risk of infection estimated from the temperature distribution image of the crowd. For example, in the item of "crowd", points according to the high risk of infection estimated from the distribution pattern in the high temperature region in the temperature distribution image of the crowd may be stored.

In the training data set 221C, the item of "retention" may store, for example, information indicating the high risk of infection estimated from a set of temperature distribution images of crowds taken at predetermined time intervals. For example, in the item of "retention", points according to the high risk of infection estimated from the distribution pattern of each high temperature region of the temperature distribution image of the crowd taken at a predetermined time interval may be stored.

In the training data set 221C, the item of "shadow" may store, for example, information indicating a high risk of infection estimated from a set of images of a human figure taken within a predetermined time. For example, in the item of "shadow", a score according to the high risk of infection estimated from the movement of a human figure included in an image taken within a predetermined time may be stored.

In the learning data set 221C, the item of "sound" may store, for example, information indicating a high risk of infection estimated from the volume information acquired within a predetermined time. For example, in the item of "sound", a score according to the high risk of infection estimated from the time-varying pattern of the volume acquired within a predetermined time may be stored.

In the training data set 221C, the item of "wind" may store, for example, information indicating a high infection risk estimated from the wind speed information acquired within a predetermined time. For example, in the item of "wind", a score according to the high risk of infection estimated from the time-varying pattern of the wind speed acquired within a predetermined time may be stored.

In the training data set 221C, the item of "teacher label" may store, for example, information indicating the high risk of infection in the environment regarding the set of sensor information identified by the data set ID. For example, the teacher label "3" indicates that it is sensor information about an environment where the risk of infection is relatively high. The teacher label "2" also indicates that the sensor information is for an environment with a moderate risk of infection. Further, the teacher label "1" indicates that the sensor information is related to the environment where the risk of infection is relatively low.

For example, the generation unit 532 can train a machine learning model based on the training data set 221C. For example, the generation unit 532 is a machine learning model based on m learning data sets (for example, m is a natural number of several thousand to tens of thousands) identified by the data sets IDs "DS-1B" to "DS-mB". Can be learned. For example, the generation unit 532 has a set of points of sensor information identified by the data set ID "DS-1B" in the machine learning model (crowd "3", retention "3", shadow "2", sound "3", When the wind "2") is input, the machine learning model may be trained so that the machine learning model outputs the teacher label "3". Further, the generation unit 532 has a set of points of sensor information identified by the data set ID "DS-2B" in the machine learning model (crowd "1", retention "1", shadow "3", sound "2", When the wind "3") is input, the machine learning model may be trained so that the machine learning model can output the teacher label "2". In this way, the generation unit 532 generates a trained machine learning model by training m training data sets identified by the data sets IDs "DS-1B" to "DS-mB". good.

(Distribution unit 533)
The distribution unit 533 may distribute information about the machine learning model generated by the generation unit 532 to the terminal device 100.

[2-3. Information processing procedure of the generator]
Next, the information processing procedure of the generator according to the embodiment will be described with reference to FIG. FIG. 14 is a diagram showing an information processing procedure of the generator according to the embodiment. First, the acquisition unit 531 can acquire information about a predetermined environment (step S101). Next, the generation unit 532 can generate a machine learning model capable of estimating the infection risk in the surrounding environment of the terminal device 100 based on the information about the predetermined environment (step S102).

[2-4. Configuration example of terminal device]
Next, the configuration of the terminal device according to the embodiment will be described with reference to FIG. FIG. 15 is a diagram showing a configuration example of the terminal device according to the embodiment. The terminal device 100 can include a communication unit 110, a sensor unit 120, a storage unit 130, an input unit 140, an output unit 150, and a control unit 160.

(Communication unit 110)
The communication unit 110 may be realized by, for example, a NIC or the like. Then, the communication unit 110 may send and receive information to and from the generation device 50. The communication unit 110 may be connected to the network by wire or wirelessly.

(Sensor unit 120)
The sensor unit 120 may include various sensors mounted on the terminal device 100. For example, the sensor unit 120 may include a temperature sensor, an image sensor (camera), a sound sensor, and a wind speed sensor. The sensor unit 120 can detect sensor information regarding the surrounding environment of the terminal device 100. It should be noted that each of the above sensors is merely an example and is not limited. That is, the sensor unit 120 may be configured to include a part of each sensor, or may include other sensors such as a humidity sensor in addition to or in place of each sensor.

The temperature sensor may be, for example, an infrared sensor. The temperature sensor can detect the temperature of an object by measuring the infrared rays emitted from the object. The temperature sensor can detect, for example, the temperature of an object in the surrounding environment of the terminal device 100. For example, the temperature sensor may be a thermography camera. A thermography camera can measure infrared rays emitted from the surface of an object and output a temperature distribution image in which colors are assigned to the surface temperature. The temperature sensor may include, for example, an infrared sensor and a thermography camera.

The image sensor (camera) can image the surrounding environment of the terminal device 100. The image sensor may be, for example, a CCD (Charge-Coupled Device) image sensor, a CMOS (Complementary Metal Oxide Semiconductor) image sensor, or the like. The image sensor (camera) can image the surrounding environment of the terminal device 100, for example.

The sound sensor can detect sound. For example, the sound sensor may be a microphone. The sound sensor can detect sound in the surrounding environment of the terminal device 100, for example.

The wind speed sensor can detect the wind speed. For example, a wind speed sensor detects wind speed, average wind speed, maximum wind speed, temperature, wind direction, etc. by starting an application and inserting a small anemometer (propeller with wings) into the headphone terminal of a smartphone or the like. May be. The wind speed sensor can detect, for example, the wind speed in the surrounding environment of the terminal device 100. The wind speed sensor may be a separate device capable of transmitting wind speed information to a smartphone or the like by wireless communication.

(Memory unit 130)
The storage unit 130 may be realized by, for example, a semiconductor memory element such as a RAM or a flash memory, or a storage device such as a hard disk or an optical disk. The storage unit 130 can store various programs (corresponding to an example of an information processing program). For example, the storage unit 130 can store the estimation program according to the embodiment. Further, the storage unit 130 can store various data. For example, the storage unit 130 can store the sensor information acquired by the acquisition unit 161. For example, the storage unit 130 can store information about the machine learning model acquired by the acquisition unit 161. For example, the storage unit 130 can store the model data of the machine learning model acquired by the acquisition unit 161. For example, the storage unit 130 can store the image captured by the image sensor.

Here, the estimation program according to the embodiment may be a program that realizes the function of the control unit 160. For example, an estimation program according to an embodiment estimates an infection risk in a predetermined environment based on an acquisition procedure for acquiring information on a predetermined environment from a sensor capable of acquiring information on the predetermined environment and information on the predetermined environment. Using a possible machine learning model, a computer is made to perform an estimation procedure for estimating the risk of infection in the surrounding environment of the terminal device 100. In the following, an application corresponding to the estimation program according to the embodiment may be described as an infection risk estimation application.

(Input unit 140, output unit 150)
The input unit 140 may be an input device that receives various operations from the user. For example, the input unit 140 may be realized by a keyboard, a mouse, operation keys, or the like. The output unit 150 may be a display device for displaying various information. That is, the output unit 150 may have a screen. For example, the output unit 150 may be realized by a liquid crystal display or the like. When a touch panel is adopted for the terminal device 100, the input unit 140 and the output unit 150 may be integrated. Further, in the following description, the output unit 150 may be described as a screen.

(Control unit 160)
The control unit 160 may be the controller of the terminal device 100. The control unit 160 may be configured by, for example, a CPU, an MPU, or the like. The control unit 160 may execute, for example, various programs (corresponding to an example of an information processing program) stored in the storage device inside the terminal device 100 using the RAM as a work area. For example, the control unit 160 may execute the estimation program according to the embodiment by using the RAM as a work area by the CPU, MPU, or the like. Further, the control unit 160 may be realized by an integrated circuit such as an ASIC or FPGA.

The control unit 160 may include an acquisition unit 161, an estimation unit 162, and an output control unit 163. The control unit 160 can realize or execute, for example, the operation of information processing described below. The internal configuration of the control unit 160 is not limited to the configuration shown in FIG. 15, and may be any other configuration as long as it is configured to perform information processing described later.

(Acquisition unit 161)
The acquisition unit 161 can acquire information about a predetermined environment. Specifically, the acquisition unit 161 can acquire the sensor information acquired by the sensor capable of acquiring the sensor information regarding the predetermined environment as an example of the information regarding the predetermined environment. For example, the acquisition unit 161 can acquire sensor information regarding the surrounding environment of the user's terminal device 100 as an example of information regarding the surrounding environment of the user's terminal device 100. For example, the acquisition unit 161 can acquire sensor information regarding the surrounding environment of the terminal device 100 detected by the sensor unit 120 from the sensor unit 120. Further, for example, the acquisition unit 161 may acquire the sensor information from a device other than the terminal device 100 such as a database in which the sensor information acquired by the terminal device 100 is stored.

For example, the acquisition unit 161 can acquire the temperature information in the surrounding environment of the terminal device 100 acquired by the temperature sensor as an example of the sensor information. For example, for example, the acquisition unit 161 can acquire a temperature distribution image in the surrounding environment of the terminal device 100 taken by a thermography camera as an example of sensor information. Further, as an example of the sensor information, the acquisition unit 161 can acquire the temperature information in the surrounding environment of the terminal device 100 acquired by the temperature sensor at predetermined time intervals. Further, as an example of the sensor information, the acquisition unit 161 can acquire the movement of a human figure included in the image in the surrounding environment of the terminal device 100 taken by the camera within a predetermined time. Further, the acquisition unit 161 can acquire the sound information in the surrounding environment of the terminal device 100 acquired by the sound sensor as an example of the sensor information. Further, the acquisition unit 161 can acquire the wind speed information in the surrounding environment of the terminal device 100 acquired by the wind speed sensor as an example of the sensor information. The acquisition unit 161 can store the acquired sensor information in the storage unit 130.

Further, the acquisition unit 161 can acquire information on the machine learning model capable of estimating the infection risk in the predetermined environment from the generation device 50 based on the sensor information on the predetermined environment. For example, the acquisition unit 161 can acquire model data of a machine learning model as an example of information regarding the machine learning model. The acquisition unit 161 can store the acquired information about the machine learning model in the storage unit 130.

(Estimation unit 162)
The estimation unit 162 may estimate the infection risk in the surrounding environment of the terminal device 100 by using a machine learning model capable of estimating the infection risk in the predetermined environment based on the sensor information regarding the predetermined environment. Specifically, the estimation unit 162 may estimate the infection risk in the surrounding environment of the terminal device 100 by using, for example, the machine learning model acquired by the acquisition unit 161. More specifically, the estimation unit 162 uses the machine learning model acquired by the acquisition unit 161 as input data for the sensor information regarding the surrounding environment of the user's terminal device 100 acquired by the acquisition unit 161 to use the terminal device 100. The risk of infection in the surrounding environment may be estimated. According to this, the terminal device 100 can estimate the infection risk at the user's current position.

(Output control unit 163)
The output control unit 163 can output the estimation result by the estimation unit 162 to the output unit 150. For example, the output control unit 163 may output the estimation result by the estimation unit 162 by voice. Further, the output control unit 163 may display the estimation result by the estimation unit 162 on the screen. For example, the output control unit 163 can notify the user of the infection risk estimated by the estimation unit 162. For example, the output control unit 163 can notify the user of the infection risk at the user's current position. For example, the output control unit 163 determines that the infection risk is high for the user when the user stays in a place estimated by the estimation unit 162 to be an environment with a high infection risk for a predetermined time or longer. You can be notified by an alarm.

The output control unit 163 may output by voice that the risk of infection is high for the user. Further, the output control unit 163 may display on the screen that the risk of infection is high for the user. For example, the output control unit 163 may select whether to output the estimation result by the estimation unit 162 by voice or display it on the screen according to the current position of the user. For example, when the output control unit 163 estimates that the user is in a public place (for example, at work or in a train) from the user's current position, the output control unit 163 displays the estimation result by the estimation unit 162 on the screen. You may. Further, when the user is estimated to be in a private place (for example, home or park) from the current position of the user, the output control unit 163 may output the estimation result by the estimation unit 162 by voice. .. Further, the output control unit 163 may share the estimation result by each terminal device 100 on the cloud to generate a risk degree map. The output control unit 163 may distribute the generated risk map to each terminal device 100. Then, the output control unit 163 may display the risk map received by each terminal device 100 on the screen of each terminal device 100.

[2-5. Example of information processing of terminal equipment]
Next, information processing of the terminal device according to the embodiment will be described with reference to FIG. FIG. 16 is a diagram for explaining an example of information processing of the terminal device according to the embodiment. In the example shown on the upper side of FIG. ₁₆ , the infection risk estimation application is installed in the terminal device 1001 of the employee of the store A. The terminal device 100 ₁ has, for example, an infection risk in the peripheral environment of the terminal device 100 ₁ (that is, an infection risk in the store A) based on sensor information regarding the surrounding environment of the terminal device 100 ₁ (that is, the environment in the store A). ) Can be estimated. For example, the terminal device 100 ₁ can be estimated to have a relatively low risk of infection in the store A. Subsequently, the terminal device 100 ₁ can upload the estimation result to the cloud. The terminal device 100 ₁ generates a risk map M1 in which a mark visually indicating that the risk of infection in the store A is relatively low (a dotted circle in the example shown in FIG. 16) is superimposed on the position of the store A. can do. Subsequently, the terminal device 100 ₁ can upload the generated risk map M1 on the HP or SNS of its own store and share it with the customer. Further, the terminal device 100 ₁ may distribute the generated risk map M1 to the terminal device 100 of another user. Then, when the terminal device 100 of another user receives the risk map M1 from the terminal device 100 ₁ , the risk map M1 can be displayed on the screen.

Further, in the example shown at the lower side of FIG. 16, the infection risk estimation application is installed in the terminal device 100 ₂ of the employee of the store B. The terminal device 100 ₂ has, for example, an infection risk in the peripheral environment of the terminal device 100 ₂ (that is, an infection risk in the store B) based on sensor information regarding the surrounding environment of the terminal device 100 ₂ (that is, the environment in the store B). ) Can be estimated. For example, the terminal device 100 ₂ can be estimated to have a relatively high risk of infection in the store B. Subsequently, the terminal device 100 ₂ can upload the estimation result to the cloud. The terminal device 100 ₂ has, for example, a risk map M1 in which a mark visually indicating that the risk of infection in the store B is relatively high (a solid circle in the example shown in FIG. 16) is superimposed on the position of the store B. Can be generated. Subsequently, the terminal device 100 ₂ can upload the generated risk map M1 on the HP or SNS of its own store and share it with the customer. Further, the terminal device 100 ₂ may distribute the generated risk map M1 to the terminal device 100 of another user. Then, when the terminal device 100 of another user receives the risk map M1 from the terminal device 100 ₁ , the risk map M1 can be displayed on the screen.

In this way, the terminal device 100 can upload the infection risk information in each store on the HP or SNS of its own store and share it with the customer. As a result, according to the terminal device 100 according to the embodiment, for example, the employee informs the customer that the inside of the store is not crowded and is safe, or informs that the inside of the store is crowded and dangerous. be able to. Further, for example, the customer can take measures such as visiting the customer at a time lag or visiting another store first. That is, according to the terminal device 100 according to the embodiment, it is possible to provide information necessary for a lifestyle in consideration of the risk of infectious diseases.

Next, information processing of the terminal device 100 according to the embodiment will be described with reference to FIG. FIG. 17 is a diagram for explaining an example of information processing of the terminal device 100 according to the embodiment. In the example shown in FIG. 17, the terminal device 100 can generate a wider range of risk map M2 based on the estimation result in the terminal device 100 of each user. When the terminal device 100 generates the risk map M2, the terminal device 100 can display the generated risk map M2 on the screen. Therefore, according to the terminal device 100 according to the embodiment, the user can consider the action schedule in consideration of the risk of infectious disease.

Next, information processing of the terminal device according to the embodiment will be described with reference to FIG. FIG. 18 is a diagram for explaining an example of information processing of the terminal device according to the embodiment. In the example shown in FIG. 18, the terminal device 100 can propose a movement route from the departure point to the destination to the user based on the risk degree map shown in FIG. For example, the terminal device 100 receives input of a departure point P1 and an arrival point P2 from a user. Subsequently, the terminal device 100 can generate a risk map M3 in which the movement route from the departure point P1 to the arrival point P2 is superimposed based on the risk map of the area including the departure point P1 and the arrival point P2. .. For example, the terminal device 100 can generate a risk map M3 in which a plurality of movement routes according to the infection risk are superimposed. In the example shown in FIG. 18, it is possible to generate a risk map M3 in which a movement route RT1 having a relatively low infection risk, a movement route RT2 having a medium infection risk, and a movement route RT3 having a relatively high infection risk are superimposed. can. Further, when the terminal device 100 generates the risk map M3, the generated risk map M3 can be displayed on the screen. Therefore, according to the terminal device 100 according to the embodiment, the user can move along a route having a relatively low risk of infectious disease.

[2-6. Information processing procedure of terminal device]
Next, the information processing procedure of the terminal device according to the embodiment will be described with reference to FIG. FIG. 19 is a diagram showing an information processing procedure of the terminal device according to the embodiment. First, the acquisition unit 161 can acquire information about a predetermined environment (step S201). Next, the estimation unit 162 can estimate the infection risk in the surrounding environment of the terminal device 100 by using a machine learning model that can estimate the infection risk in the predetermined environment based on the information about the predetermined environment (step). S202). Then, the output control unit 163 can output the estimation result by the estimation unit 162 (step S203).

[3. effect〕
As described above, the terminal device 100 according to the embodiment includes an acquisition unit 161 and an estimation unit 162. The acquisition unit 161 acquires information on a predetermined environment from a sensor that can acquire information on the predetermined environment. The estimation unit 162 estimates the infection risk in the surrounding environment of the terminal device 100 by using a machine learning model that can estimate the infection risk in the predetermined environment based on the information about the predetermined environment.

Thereby, the terminal device 100 can estimate the infection risk in the surrounding environment of the terminal device 100 by using a machine learning model having a high correlation of various sensor information.

Further, the information regarding the predetermined environment includes a temperature distribution image in the surrounding environment of the terminal device 100, a temperature distribution image in the peripheral environment of the terminal device 100 acquired at a predetermined time interval, and a periphery of the terminal device 100 taken within a predetermined time. It includes an image of a human figure in the environment, sound information in the surrounding environment of the terminal device 100, and at least one of wind velocity information in the surrounding environment of the terminal device 100.

Thereby, the terminal device 100 can estimate the infection risk in the surrounding environment of the terminal device 100 based on various sensor information.

Further, the terminal device 100 according to the embodiment further includes an output control unit 163 that outputs an estimation result by the estimation unit 162 to the terminal device 100.

As a result, the terminal device 100 can provide the user of the terminal device 100 with information necessary for a lifestyle considering the risk of infectious diseases.

Further, the output control unit 163 displays a risk map showing the infection risk based on the estimation result estimated by the estimation unit 162 on the screen of the user's terminal device 100.

This enables the terminal device 100 to allow the user of the terminal device 100 to consider the action schedule in consideration of the risk of infectious disease.

Further, the output control unit 163 displays the movement route from the departure point designated by the user to the destination on the screen of the user's terminal device 100 based on the estimation result estimated by the estimation unit 162.

Thereby, the terminal device 100 enables the user of the terminal device 100 to move along a route having a relatively low risk of infectious disease.

Further, the generation device 50 according to the embodiment includes an acquisition unit 531 and a generation unit 532. The acquisition unit 531 acquires information about a predetermined environment. The generation unit 532 generates a machine learning model capable of estimating the infection risk in the surrounding environment of the terminal device 100 based on the information about the predetermined environment.

As a result, the generator 50 can estimate the risk of being infected with an infectious disease by using a machine learning model with high correlation of various sensor information.

Further, the acquisition unit 531 acquires the sensor information acquired by the sensor capable of acquiring the sensor information regarding the predetermined environment as the information regarding the predetermined environment. The generation unit 532 has learned that when sensor information is input to the machine learning model, the sensor information is classified into at least one of two or more clusters according to the high risk of infection. Generate a model.

Thereby, for example, the generator 50 can estimate the risk of being infected with an infectious disease even when the level of high risk of infection is two types, "danger" and "OK". ..

Further, when the sensor information is input as the input information of the machine learning model, the generation unit 532 outputs the label indicating the infection risk corresponding to the cluster in which the sensor information is classified as the output information of the machine learning model. Generate a trained machine learning model.

This makes it possible for the generation device 50 to output an estimation result that is easy for the user to understand.

Further, the acquisition unit 531 acquires sensor information which is a temperature distribution image in the surrounding environment of the terminal device 100. The generation unit 532 generates a machine learning model capable of estimating the infection risk in the surrounding environment of the terminal device 100 based on the temperature distribution pattern of the temperature distribution image.

Thereby, the generation device 50 can estimate the infection risk, for example, based on the information indicating the presence or absence of crowds.

Further, the acquisition unit 531 acquires sensor information which is a temperature distribution image in the surrounding environment of the terminal device 100 acquired at predetermined time intervals. The generation unit 532 generates a machine learning model capable of estimating the infection risk in the surrounding environment of the terminal device 100 based on the temperature distribution pattern of each temperature distribution image.

Thereby, the generation device 50 can estimate the infection risk, for example, based on the information indicating the presence or absence of human retention.

Further, the acquisition unit 531 acquires sensor information which is an image of a human figure in the surrounding environment of the terminal device 100 taken within a predetermined time. The generation unit 532 generates a machine learning model capable of estimating the infection risk in the surrounding environment of the terminal device 100 based on the time-varying pattern of the human figure included in the image.

As a result, the generation device 50 can estimate the risk of infection while considering privacy.

Further, the acquisition unit 531 acquires sensor information which is sound information in the surrounding environment of the terminal device 100 acquired within a predetermined time. The generation unit 532 generates a machine learning model capable of estimating the infection risk in the surrounding environment of the terminal device 100 based on the pattern of the time change of the sound information.

Thereby, the generation device 50 can estimate the infection risk based on the information on the movement of the person estimated from the sound information, for example.

Further, the acquisition unit 531 acquires sensor information which is wind speed information in the surrounding environment of the terminal device 100 acquired within a predetermined time. The generation unit 532 generates a machine learning model capable of estimating the infection risk in the surrounding environment of the terminal device 100 based on the pattern of the time change of the wind speed information.

This makes it possible for the generator 50 to estimate the risk of infection, for example, based on information such as whether the air circulation is good or bad and whether the ventilation is sufficient or insufficient.

[4. Hardware configuration]
Further, the terminal device 100 or the generation device 50 according to the above-described embodiment may be realized by, for example, a computer 1000 having a configuration as shown in FIG. FIG. 20 is a hardware configuration diagram showing an example of a computer that realizes the functions of the terminal device 100 or the generation device 50. The computer 1000 may include a CPU 1100, a RAM 1200, a ROM 1300, an HDD 1400, a communication interface (I / F) 1500, an input / output interface (I / F) 1600, and a media interface (I / F) 1700.

The CPU 1100 operates based on a program stored in the ROM 1300 or the HDD 1400, and controls each part. The ROM 1300 can store a boot program executed by the CPU 1100 when the computer 1000 is started, a program depending on the hardware of the computer 1000, and the like.

The HDD 1400 can store a program executed by the CPU 1100, data used by such a program, and the like. The communication interface 1500 can receive data from another device via a predetermined communication network and send the data to the CPU 1100, and can transmit the data generated by the CPU 1100 to the other device via the predetermined communication network.

The CPU 1100 can control an output device such as a display or a printer and an input device such as a keyboard or a mouse via the input / output interface 1600. The CPU 1100 can acquire data from the input device via the input / output interface 1600. Further, the CPU 1100 can output the generated data to the output device via the input / output interface 1600.

The media interface 1700 can read a program or data stored in the recording medium 1800 and provide the program or data to the CPU 1100 via the RAM 1200. The CPU 1100 can load such a program from the recording medium 1800 onto the RAM 1200 via the media interface 1700 and execute the loaded program. The recording medium 1800 is, for example, an optical recording medium such as a DVD (Digital Versatile Disc) or PD (Phase change rewritable Disk), a magneto-optical recording medium such as MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory. And so on.

For example, when the computer 1000 functions as the terminal device 100 or the generation device 50 according to the embodiment, the CPU 1100 of the computer 1000 executes the function of the control unit 160 or the control unit 53 by executing the program loaded on the RAM 1200. It can be realized. The CPU 1100 of the computer 1000 reads these programs from the recording medium 1800 and executes them, but as another example, these programs may be acquired from another device via a predetermined communication network.

Although some of the embodiments of the present application have been described in detail with reference to the drawings, these are examples, and various modifications are made based on the knowledge of those skilled in the art, including the embodiments described in the disclosure column of the invention. It is possible that the present invention can be practiced in other modified forms.

[5. others〕
Further, among the processes described in the above-described embodiments and modifications, all or part of the processes described as being automatically performed can be manually performed, or are described as being manually performed. It is also possible to automatically perform all or part of the processed processing by a known method. In addition, information including processing procedures, specific names, various data and parameters shown in the above documents and drawings can be arbitrarily changed unless otherwise specified. For example, the various information shown in each figure is not limited to the information shown in the figure.

Further, each component of each of the illustrated devices is a functional concept, and does not necessarily have to be physically configured as shown in the figure. That is, the specific form of distribution / integration of each device is not limited to the one shown in the figure, and all or part of them may be functionally or physically distributed / physically distributed in any unit according to various loads and usage conditions. Can be integrated and configured.

Further, the above-described embodiments and modifications can be appropriately combined as long as the processing contents do not contradict each other.

Further, the above-mentioned "section, module, unit" can be read as "means" or "circuit". For example, the estimation unit can be read as an estimation means or an estimation circuit.

1 Information processing system 50 Generation device 51 Communication section 52 Storage section 53 Control section 531 Acquisition section 532 Generation section 533 Distribution section 100 Terminal device 110 Communication section 120 Sensor section 130 Storage section 140 Input section 150 Output section 160 Control section 161 Acquisition section 162 Estimator 163 Output control

Claims (16)

An acquisition procedure for acquiring information on a predetermined environment from a sensor capable of acquiring information on the predetermined environment, and an acquisition procedure.
An estimation procedure for estimating the infection risk in the surrounding environment of the terminal device using a machine learning model capable of estimating the infection risk in the predetermined environment based on the information on the predetermined environment, and an estimation procedure.
Let the computer run
The acquisition procedure is
As information on the predetermined environment, a plurality of temperature distribution images in the surrounding environment of the terminal device taken at predetermined time intervals by a thermography camera mounted on the terminal device are acquired.
The estimation procedure is
It is possible to estimate the infection risk in the surrounding environment of the terminal device based on the information showing the time change of the distribution pattern in the high temperature region, which is the region above the predetermined temperature in each of the plurality of temperature distribution images taken at the predetermined time interval. The risk of infection in the surrounding environment of the terminal device is estimated from the information showing the time change of the distribution pattern in the high temperature region in each of the plurality of temperature distribution images taken at the predetermined time interval by using the machine learning model. do,
Estimate program. Information about the given environment
A temperature distribution image in the surrounding environment of the terminal device, a temperature distribution image in the peripheral environment of the terminal device acquired at predetermined time intervals, an image of a human figure in the peripheral environment of the terminal device taken within a predetermined time, and the terminal device. Includes sound information in the surrounding environment of the terminal device and at least one of wind velocity information in the surrounding environment of the terminal device.
The estimation program according to claim 1. An output control procedure that outputs the estimation result by the estimation procedure to the terminal device,
The estimation program according to claim 1 or 2, further causing a computer to execute the program. The output control procedure is
This is a procedure for displaying a risk map showing an infection risk based on the estimation result estimated by the estimation procedure on the screen of the terminal device.
The estimation program according to claim 3. The output control procedure is
Based on the estimation result estimated by the estimation procedure, the procedure is to display the movement route from the departure point to the destination specified by the user of the terminal device on the screen of the terminal device.
The estimation program according to claim 3 or 4. An acquisition unit that acquires information on the predetermined environment from a sensor that can acquire information on the predetermined environment, and an acquisition unit.
An estimation unit that estimates the infection risk in the surrounding environment of the terminal device using a machine learning model that can estimate the infection risk in the predetermined environment based on the information about the predetermined environment.
Equipped with
The acquisition unit
As information on the predetermined environment, a plurality of temperature distribution images in the surrounding environment of the terminal device taken at predetermined time intervals by a thermography camera mounted on the terminal device are acquired.
The estimation unit is
It is possible to estimate the infection risk in the surrounding environment of the terminal device based on the information showing the time change of the distribution pattern in the high temperature region, which is the region above the predetermined temperature in each of the plurality of temperature distribution images taken at the predetermined time interval. The risk of infection in the surrounding environment of the terminal device is estimated from the information showing the time change of the distribution pattern in the high temperature region in each of the plurality of temperature distribution images taken at the predetermined time interval by using the machine learning model. do,
Estimator. The acquisition step of acquiring the information on the predetermined environment from the sensor capable of acquiring the information on the predetermined environment, and the acquisition step.
Using a machine learning model that can estimate the infection risk in the predetermined environment based on the information about the predetermined environment, the estimation step for estimating the infection risk in the surrounding environment of the terminal device, and the estimation step.
Including
The acquisition step is
As information on the predetermined environment, a plurality of temperature distribution images in the surrounding environment of the terminal device taken at predetermined time intervals by a thermography camera mounted on the terminal device are acquired.
The estimation step is
It is possible to estimate the infection risk in the surrounding environment of the terminal device based on the information showing the time change of the distribution pattern in the high temperature region, which is the region above the predetermined temperature in each of the plurality of temperature distribution images taken at the predetermined time interval. The risk of infection in the surrounding environment of the terminal device is estimated from the information showing the time change of the distribution pattern in the high temperature region in each of the plurality of temperature distribution images taken at the predetermined time interval by using the machine learning model. do,
Estimating method. An acquisition procedure for acquiring information on a predetermined environment from a sensor capable of acquiring information on the predetermined environment, and an acquisition procedure.
A generation procedure for generating a machine learning model capable of estimating the infection risk in the surrounding environment of the terminal device based on the information on the predetermined environment, and a generation procedure.
Let the computer run
The acquisition procedure is
As information on the predetermined environment, a plurality of temperature distribution images in the surrounding environment of the terminal device taken at predetermined time intervals by a thermography camera mounted on the terminal device are acquired.
The generation procedure is
It is possible to estimate the infection risk in the surrounding environment of the terminal device based on the information showing the time change of the distribution pattern in the high temperature region, which is the region above the predetermined temperature in each of the plurality of temperature distribution images taken at the predetermined time interval. Generate the machine learning model,
Generation program. The acquisition procedure is
As the information regarding the predetermined environment, it is a procedure for acquiring the sensor information acquired by a sensor capable of acquiring the sensor information regarding the predetermined environment.
The generation procedure is
When the sensor information is input to the machine learning model, the machine learning model learned so that the sensor information is classified into at least one of two or more clusters according to the high risk of infection. Is the procedure to generate
The generation program according to claim 8. The generation procedure is
When the sensor information is input as the input information of the machine learning model, the machine learning model is learned to output a label indicating the infection risk corresponding to the cluster in which the sensor information is classified as the output information of the machine learning model. The procedure for generating a machine learning model,
The generation program according to claim 9. The acquisition procedure is
This is a procedure for acquiring the sensor information which is a temperature distribution image in the surrounding environment of the terminal device.
The generation procedure is
This is a procedure for generating a machine learning model capable of estimating the infection risk in the surrounding environment of the terminal device based on the temperature distribution pattern of the temperature distribution image.
The generation program according to claim 9. The acquisition procedure is
It is a procedure for acquiring the sensor information which is an image of a human figure in the surrounding environment of the terminal device taken within a predetermined time.
The generation procedure is
It is a procedure for generating a machine learning model capable of estimating the infection risk in the surrounding environment of the terminal device based on the time-varying pattern of the human figure included in the image.
The generation program according to any one of claims 9 to 11. The acquisition procedure is
It is a procedure for acquiring the sensor information which is the sound information in the surrounding environment of the terminal device acquired within a predetermined time.
The generation procedure is
This is a procedure for generating a machine learning model capable of estimating the infection risk in the surrounding environment of the terminal device based on the pattern of time change of the sound information.
The generation program according to any one of claims 9 to 12. The acquisition procedure is
It is a procedure for acquiring the sensor information which is the wind speed information in the surrounding environment of the terminal device acquired within a predetermined time.
The generation procedure is
It is a procedure for generating a machine learning model capable of estimating the infection risk in the surrounding environment of the terminal device based on the pattern of the time change of the wind speed information.
The generation program according to any one of claims 9 to 13. An acquisition unit that acquires information on the predetermined environment from a sensor that can acquire information on the predetermined environment, and an acquisition unit.
A generator that generates a machine learning model that can estimate the risk of infection in the surrounding environment of the terminal device based on the information about the predetermined environment.
Equipped with
The acquisition unit
As information on the predetermined environment, a plurality of temperature distribution images in the surrounding environment of the terminal device taken at predetermined time intervals by a thermography camera mounted on the terminal device are acquired.
The generator is
It is possible to estimate the infection risk in the surrounding environment of the terminal device based on the information showing the time change of the distribution pattern in the high temperature region, which is the region above the predetermined temperature in each of the plurality of temperature distribution images taken at the predetermined time interval. Generate the machine learning model,
Generator. The acquisition step of acquiring the information on the predetermined environment from the sensor capable of acquiring the information on the predetermined environment, and the acquisition step.
A generation step to generate a machine learning model capable of estimating the infection risk in the surrounding environment of the terminal device based on the information about the predetermined environment, and
Including
The acquisition step is
As information on the predetermined environment, a plurality of temperature distribution images in the surrounding environment of the terminal device taken at predetermined time intervals by a thermography camera mounted on the terminal device are acquired.
The generation step is
It is possible to estimate the infection risk in the surrounding environment of the terminal device based on the information showing the time change of the distribution pattern in the high temperature region, which is the region above the predetermined temperature in each of the plurality of temperature distribution images taken at the predetermined time interval. Generate the machine learning model,
Generation method.

Images