Risk estimation assistance program, risk estimation assistance device, risk estimation assistance method, trained model generation method, trained model generation device, trained model generation program, and recording medium

Abstract

The present invention provides a program for estimating a risk that may occur in the future to a subject, on the basis of time-series information relating to the subject.　A risk estimation assistance program according to the present disclosure includes an information acquisition procedure, a risk calculation procedure, and a risk output procedure, and is for causing a computer to execute the procedures. The information acquisition procedure acquires time-series information indicating a feature of a subject, the risk calculation procedure calculates a risk that may occur in the future to the subject on the basis of the time-series information, and the risk output procedure outputs the risk.

Description

Risk estimation support program, risk estimation support device, risk estimation support method, trained model generation method, trained model generation device, trained model generation program, and recording medium

[0001] The present disclosure relates to a risk estimation support program, a risk estimation support device, a risk estimation support method, a trained model generation method, a trained model generation device, a trained model generation program, and a recording medium.

[0002] Patent Document 1 discloses a method for obtaining information about diabetes in a subject, which includes measuring the sterol uptake capacity of lipoproteins in a biological sample from the subject, and the measured value of the sterol uptake capacity per unit volume of the biological sample serves as an indicator of the risk of developing diabetes.

[0003] Japanese Patent Application Laid-Open No. 2020-051928

[0004] In the risk estimation method disclosed in Patent Document 1, risk is estimated from measurements such as current test values ​​of a subject. However, when estimating a subject's future risk, it is important to take into account not only the subject's current condition but also the subject's past condition and past trends. By taking into account past conditions, etc., it is believed that more accurate risk estimation can be achieved. This applies not only to the disease risk estimation disclosed in Patent Document 1, but also to various risk estimations for any subject.

[0005] Therefore, the present disclosure aims to provide a risk estimation support program, a risk estimation support device, a risk estimation support method, a trained model generation method, a trained model generation device, a trained model generation program, and a recording medium for estimating future risks that may occur to a target based on the target's time series information.

[0006] In order to achieve the above-mentioned objective, the risk estimation support program disclosed herein includes an information acquisition procedure, a risk calculation procedure, and a risk output procedure, wherein the information acquisition procedure acquires time-series information indicating the characteristics of the object, the risk calculation procedure calculates future risks that may occur to the object based on the time-series information, and the risk output procedure outputs the risks.The program is for causing a computer to execute each of these procedures.

[0007] The trained model generation program of the present disclosure includes a training data acquisition procedure and a trained model generation procedure, wherein the training data acquisition procedure acquires time series information indicating characteristics of a target as training data, and the trained model generation procedure generates a trained model for outputting future risks to the target from the time series information through machine learning using the training data, and is a program for causing a computer to execute each of the above procedures.

[0008] The trained model of the present disclosure is machine-learned using features, which represent the time series trends of an object based on an arbitrary function converted from the object's time series information, and is a trained model for causing a computer to function so as to output future risks that may occur to the object from the time series information.

[0009] The risk estimation support device disclosed herein includes an information acquisition unit, a risk calculation unit, and a risk output unit, wherein the information acquisition unit acquires time-series information indicating characteristics of an object, the risk calculation unit calculates future risks that may occur to the object based on the time-series information, and the risk output unit outputs the risks.

[0010] The trained model generation device of the present disclosure includes a training data acquisition unit and a trained model generation unit, wherein the training data acquisition unit acquires time series information indicating characteristics of a target as training data, and the trained model generation unit generates a trained model for outputting future risks that may occur to the target from the time series information through machine learning using the training data.

[0011] The risk estimation support method disclosed herein includes an information acquisition step, a risk calculation step, and a risk output step, wherein the information acquisition step acquires time-series information indicating characteristics of an object, the risk calculation step calculates a risk that may occur to the object in the future based on the time-series information, and the risk output step outputs the risk, and each of the steps is executed by a computer.

[0012] The trained model generation method of the present disclosure includes a training data acquisition step and a trained model generation step, wherein the training data acquisition step acquires time-series information indicating characteristics of a target as training data, and the trained model generation step generates a trained model for outputting future risks to the target from the time-series information through machine learning using the training data, and each of the steps is executed by a computer.

[0013] The recording medium of the present disclosure is a computer-readable recording medium on which the program of the present disclosure is recorded.

[0014] According to the present disclosure, it is possible to estimate possible future risks based on time-series information of a target.

[0015] FIG. 1 is a block diagram showing the configuration of an example of a risk estimation support device disclosed herein. FIG. 2 is a block diagram showing an example of the hardware configuration of the risk estimation support device disclosed herein. FIG. 3 is a flowchart showing an example of procedures in a risk estimation support program disclosed herein. FIG. 4 is a block diagram showing the configuration of an example of a trained model generation device disclosed herein. FIG. 5 is a block diagram showing an example of the hardware configuration of a trained model generation device disclosed herein. FIG. 6 is a flowchart showing an example of procedures in a trained model generation program disclosed herein. FIG. 7 is a block diagram showing the configuration of another example of a trained model generation device disclosed herein. FIG. 8 is a flowchart showing another example of procedures in a trained model generation program disclosed herein. FIG. 9 is a diagram showing an example of converting time-series information into a function and a feature space. FIG. 10 is a diagram showing an example of a procedure for creating a trained model for estimating risk.

[0016] An embodiment of the present disclosure will be described. Note that the present disclosure is not limited to the following embodiment. Note that in the following drawings, the same parts are assigned the same reference numerals. Furthermore, the descriptions of the embodiments can be used interchangeably unless otherwise specified. Furthermore, the configurations of the embodiments can be combined unless otherwise specified. Furthermore, in each procedure described below in the program of the present disclosure, for example, "procedure" can be read as "processing."

[0017] It should be noted that the "risk" in the present disclosure is not particularly limited, but examples thereof include disease risk, health risk, safety risk, accident risk, breakdown risk, disaster risk, and economic risk.

[0018] First Embodiment A risk estimation support program, a risk estimation support device, and a risk estimation support method according to the present disclosure will be described.

[0019] The risk estimation support program disclosed herein is a program for causing a computer to execute an information acquisition procedure, a risk calculation procedure, and a risk output procedure. The risk estimation support program disclosed herein can also be said to be a program for causing a computer to function as the information acquisition procedure, the risk calculation procedure, and the risk output procedure. Furthermore, the risk estimation support program disclosed herein can also be said to be a program for causing a computer to execute, for example, each step of a risk estimation support method described below.

[0020] Next, an example of a risk estimation support device according to the present disclosure will be described with reference to FIGS.

[0021] 1 is a block diagram showing an example of the configuration of a risk estimation support device 10 (the device 10) according to the present disclosure. As shown in FIG. 1, the device 10 includes an information acquisition unit 11, a risk calculation unit 12, and a risk output unit 13.

[0022] The device 10 may be, for example, a single device including the above-mentioned components, or a device in which the components can be connected via a communication network. The device 10 may also be connected to an external device (described later) via the communication network. The communication network is not particularly limited and may be any known network, for example, wired or wireless. Examples of the communication network include the Internet, the World Wide Web (WWW), a telephone line, a Local Area Network (LAN), a Storage Area Network (SAN), a Delay Tolerant Networking (DTN), a Low Power Wide Area Network (LPWA), and a Local 5G (L5G). Examples of wireless communication include Wi-Fi (registered trademark), Bluetooth (registered trademark), local 5G, and LPWA. The wireless communication may be a form in which each device communicates directly (ad hoc communication), infrastructure communication, indirect communication via an access point, or the like. The device 10 may be incorporated into a server as a system. The device 10 may also be, for example, a personal computer (PC, e.g., desktop or laptop) on which the program of the present disclosure is installed, a smartphone, a tablet terminal, digital signage, or the like. The device 10 may be, for example, in the form of cloud computing or edge computing, in which at least one of the components is located on a server and the other components are located on a terminal.

[0023] 2 is a block diagram illustrating an example of the hardware configuration of the device 10. The device 10 includes, for example, a central processing unit (CPU, GPU, etc.) 101, a memory 102, a bus 103, a storage device 104, an input device 105, an output device 106, and a communication device 107. The components of the device 10 are connected to each other via the bus 103 and their respective interfaces (I / F).

[0024] The central processing unit 101 operates in cooperation with other components via a controller (system controller, I / O controller, etc.) and the like, and is responsible for overall control of the device 10. In the device 10, the central processing unit 101 executes, for example, the program disclosed herein and other programs, and also reads and writes various types of information. Specifically, for example, the central processing unit 101 functions as an information acquisition unit 11, a risk calculation unit 12, and a risk output unit 13. The central processing unit 101 may include, as a computing device, a CPU, a GPU (Graphics Processing Unit), an APU (Accelerated Processing Unit), or a combination of these.

[0025] The bus 103 can also be connected to, for example, external devices. Examples of the external devices include external storage devices such as external databases, printers, external input devices, external display devices, and external imaging devices. The device 10 can be connected to an external network (the communication line network) by, for example, a communication device 107 connected to the bus 103, and can also be connected to other devices via the external network.

[0026] The memory 102 may be, for example, a main memory (primary storage device). When the central processing unit 101 performs processing, the memory 102 reads various operating programs, such as the program of the present disclosure, stored in the storage device 104 (described later), and the central processing unit 101 receives data from the memory 102 and executes the programs. The main memory may be, for example, a RAM (random access memory). Alternatively, the memory 102 may be, for example, a ROM (read only memory).

[0027] The storage device 104 is also referred to as an auxiliary storage device, for example, in contrast to the main memory (primary storage device). As described above, the storage device 104 stores an operating program including the program of the present disclosure. The storage device 104 may be, for example, a combination of a recording medium and a drive that reads and writes from and to the recording medium. The recording medium is not particularly limited and may be, for example, an internal or external type, and examples include a hard disk (HD), CD-ROM, CD-R, CD-RW, MO, DVD, flash memory, and memory card. The storage device 104 may be, for example, a hard disk drive (HDD) or a solid state drive (SSD) in which the recording medium and drive are integrated.

[0028] In the present device 10, the memory 102 and the storage device 104 can also store various information such as log information, information acquired from an external database (not shown) or an external device, information generated by the present device 10, and information used when the present device 10 executes processing. In this case, the memory 102 and the storage device 104 may store, for example, time-series information, possible future risks to the target, feature quantities, etc., as described below. Note that at least a portion of the information may be stored, for example, in an external server other than the memory 102 and the storage device 104, or may be stored in a distributed manner across multiple terminals using blockchain technology or the like.

[0029] The device 10 further includes, for example, an input device 105 and an output device 106. Examples of the input device 105 include pointing devices such as a touch panel, track pad, or mouse; a keyboard; imaging means such as a camera or scanner; card readers such as an IC card reader or a magnetic card reader; and audio input means such as a microphone. Examples of the output device 106 include display devices such as an LED display or a liquid crystal display; audio output devices such as a speaker; a printer; etc. In the present disclosure, the input device 105 and the output device 106 are configured separately, but the input device 105 and the output device 106 may also be configured as an integrated device, such as a touch panel display.

[0030] First, an example of the processing of the risk estimation support program of the present disclosure will be specifically described with reference to Fig. 3. Fig. 3 is a flowchart showing an example of each procedure of the risk estimation support program of the present disclosure.

[0031] First, the information acquisition unit 11 acquires time-series information indicating the characteristics of a target (S11, information acquisition procedure). Examples of the target include a person, an object, an environment, a machine, a facility, a system, and any organization. The target may be appropriately selected depending on the target for which risk is to be calculated, for example.

[0032] The time series information is, for example, time series information indicating characteristics of the target at the present and in the past. The time series information may be, for example, time series information at an arbitrary time, period, age, month, or day. The time series information may be, for example, time series information at an arbitrary number of time points. The arbitrary number of time points is not particularly limited and may be set, for example, according to the number of pieces of time series information available. The time series information may also be, for example, time series information at an arbitrary time interval. The arbitrary time interval is not particularly limited and may be set, for example, according to the time interval of the pieces of time series information available. In the present disclosure, if at least one of time series information at an arbitrary number of time points and time series information at an arbitrary time interval can be used as the time series information, the risk estimation method of the present disclosure can replace, for example, conventional risk estimation methods that can estimate risk only using time series information at a certain number of time points or time series information at a certain time interval.

[0033] Examples of the time-series information include time-series information related to health data, accident data, failure data, disaster data, and economic data. Examples of the health data include test values ​​measured by health checkups, home measuring devices, and wearable devices, lifestyle data based on questionnaire information, and information identifying the subject, such as age and gender. Examples of the lifestyle data include sleep habits, exercise habits, eating habits, smoking habits, drinking habits, stress habits, and habits related to hobbies or entertainment. Examples of the accident data include accident statistics, safety records, accident occurrence data, and environmental data at the time of the accident. Examples of the failure data include maintenance history, usage history, lifespan data, quality data, and error logs of equipment or facilities. Examples of the disaster data include seismic activity data, meteorological data, topographical and geological data, population data, and infrastructure data. Examples of the economic data include market data, credit rating data, economic indicators, financial reports, and financial data.

[0034] Next, the risk calculation unit 12 calculates a future risk that may occur to the subject based on the time-series information (S12, risk calculation procedure). The risk calculation unit 12 may, for example, convert the time-series information into an arbitrary function and calculate the risk based on the function. When the time-series information includes the health data and the health data is the results of blood tests over several years, for example, the results of the blood tests over several years can be converted into an arbitrary function and the risk can be calculated based on the function. The function is not particularly limited, and examples thereof include a linear function (linear function), a quadratic function, a cubic function, a polynomial function, an exponential function, a logarithmic function, a trigonometric function, an inverse trigonometric function, a hyperbolic function, a variance function, and a special function.

[0035] The risk calculation unit 12 may, for example, extract a feature representing a time-series trend of the target based on the function and calculate the risk based on the feature. The risk calculation unit 12 may, for example, convert the feature into a feature space representing a time-series trend and calculate the risk based on the feature space. The feature is, for example, a parameter for determining the characteristics of the function. For example, if the function is a linear function, the parameters include a slope and an intercept. In this case, for example, both the slope and the intercept may be the feature, or either the slope or the intercept may be the feature. For example, for other functions, at least one parameter for determining the characteristics of each function may be the feature.

[0036] The risk calculation unit 12 may use, for example, a trained model. That is, the risk calculation unit 12 calculates the risk using, for example, a trained model. The trained model may be, for example, a trained model of the present disclosure described below. The trained model may be, for example, a trained model that is machine-learned using feature quantities, and the feature quantities represent a time-series trend of the target based on an arbitrary function converted from time-series information of the target. The trained model may be, for example, a model that performs survival analysis such as Cox regression analysis using the feature quantities as covariates.

[0037] Next, the risk output unit 13 outputs the risk (S13, risk output procedure). The output is not particularly limited, and may be output by the output device 106 of the device 10, for example, or may be output by an output means provided in a device other than the device 10. The output format of the risk is not particularly limited, and examples thereof include text, a matrix, a heat map, a graph, a radar chart, a tree, a table, a dashboard, and a histogram.

[0038] According to the risk estimation support program of the present disclosure, the information acquisition step acquires time-series information indicating the characteristics of the target, the risk calculation step calculates a future risk that may occur to the target based on the time-series information, and the risk output step outputs the risk. This makes it possible to estimate the future risk that may occur to the target based on the time-series information of the target. Furthermore, even if the time-series information is, for example, at least one of time-series information consisting of an arbitrary number of time points and time-series information at an arbitrary time interval, the risk can be estimated. This means that, for example, even in cases where risk prediction was previously impossible due to missing values ​​or when risk prediction is desired using data with no periodicity, the risk estimation support program of the present disclosure can estimate the risk.

[0039] Next, the risk estimation support method of the present disclosure will be described.

[0040] The risk estimation support method of the present disclosure can incorporate the descriptions of the risk estimation support program and risk estimation support device of the present disclosure. The risk estimation support method of the present disclosure is, for example, a method implemented by replacing each "procedure" in the program of the present disclosure with a "step." Specifically, the risk estimation support method of the present disclosure includes an information acquisition step, a risk calculation step, and a risk output step. The risk estimation support method of the present disclosure can be implemented using, for example, the device 10 of FIG. 1. Note that the risk estimation support method of the present disclosure is not limited to use with the device 10 of FIG. 1.

[0041] [Embodiment 2] Next, a trained model generation program, a trained model generation device, a trained model generation method, and a trained model according to the present disclosure will be described.

[0042] The trained model generation program of the present disclosure is a program for causing a computer to execute a training data acquisition procedure and a trained model generation procedure. The trained model generation program of the present disclosure can also be said to be a program for causing a computer to function as the training data acquisition procedure and the trained model generation procedure. Furthermore, the trained model generation program of the present disclosure can also be said to be a program for causing a computer to execute, for example, each step of a trained model generation method described below.

[0043] Next, an example of a trained model generation device according to the present disclosure will be described with reference to FIGS. 4 and 5.

[0044] 4 is a block diagram showing an example configuration of a trained model generation device 20 (the present device 20) according to the present disclosure. As shown in FIG. 4 , the present device 20 includes a training data acquisition unit 21 and a trained model generation unit 22.

[0045] The device 20 may be, for example, a single device including the above-mentioned units, or a device in which the above-mentioned units can be connected via a communication network. The device 20 may also be connected to an external device (described later) via the communication network. The communication network is not particularly limited and may be a known network, for example, wired or wireless. Examples of the communication network include the Internet, the World Wide Web (WWW), a telephone line, a Local Area Network (LAN), a Storage Area Network (SAN), a Delay Tolerant Networking (DTN), a Low Power Wide Area Network (LPWA), and a Local 5G (L5G). Examples of wireless communication include Wi-Fi (registered trademark), Bluetooth (registered trademark), local 5G, and LPWA. The wireless communication may be a form in which each device communicates directly (ad hoc communication), infrastructure communication, indirect communication via an access point, or the like. The device 20 may be incorporated into a server as a system. The device 20 may also be, for example, a personal computer (PC, e.g., desktop or notebook type) on which the program of the present disclosure is installed, a smartphone, a tablet terminal, digital signage, or the like. The device 20 may be, for example, in the form of cloud computing or edge computing, in which at least one of the units is located on a server and the other units are located on a terminal.

[0046] 5 shows a block diagram of the hardware configuration of the device 20. The device 20 includes, for example, a central processing unit (CPU, GPU, etc.) 201, a memory 202, a bus 203, a storage device 204, an input device 205, an output device 206, and a communication device 207. The components of the device 20 are connected to each other via the bus 203 and their respective interfaces (I / F).

[0047] The central processing unit 201 operates in cooperation with other components via a controller (such as a system controller or an I / O controller) and is responsible for overall control of the device 20. In the device 20, the central processing unit 201 executes, for example, the program disclosed herein and other programs, and also reads and writes various types of information. Specifically, for example, the central processing unit 201 functions as a training data acquisition unit 21 and a trained model generation unit 22. The central processing unit 201 may include, as a computing device, a CPU, a GPU (Graphics Processing Unit), an APU (Accelerated Processing Unit), or a combination thereof.

[0048] The bus 203 can also be connected to, for example, external devices. Examples of the external devices include external storage devices such as external databases, printers, external input devices, external display devices, and external imaging devices. The device 20 can be connected to an external network (the communication line network) by, for example, a communication device 207 connected to the bus 203, and can also be connected to other devices via the external network.

[0049] The memory 202 may be, for example, a main memory (primary storage device). When the central processing unit 201 performs processing, the memory 202 reads various operating programs, such as the program of the present disclosure, stored in the storage device 204 (described later), and the central processing unit 201 receives data from the memory 202 and executes the programs. The main memory may be, for example, a RAM (random access memory). Alternatively, the memory 202 may be, for example, a ROM (read only memory).

[0050] The storage device 204 is also referred to as an auxiliary storage device, for example, in contrast to the main memory (primary storage device). As described above, the storage device 204 stores an operating program including the program of the present disclosure. The storage device 204 may be, for example, a combination of a recording medium and a drive that reads and writes from and to the recording medium. The recording medium is not particularly limited and may be, for example, an internal or external type, and examples include a hard disk (HD), CD-ROM, CD-R, CD-RW, MO, DVD, flash memory, and memory card. The storage device 204 may be, for example, a hard disk drive (HDD) or a solid state drive (SSD) in which the recording medium and drive are integrated.

[0051] In the present device 20, the memory 202 and the storage device 204 can also store various information such as log information, information acquired from an external database (not shown) or an external device, information generated by the present device 20, and information used when the present device 20 executes processing. In this case, the memory 202 and the storage device 204 may store, for example, time-series information, feature quantities, etc., which will be described later. Note that at least a portion of the information may be stored, for example, in an external server other than the memory 202 and the storage device 204, or may be stored in a distributed manner across multiple terminals using blockchain technology or the like.

[0052] The device 20 further includes, for example, an input device 205 and an output device 206. Examples of the input device 205 include pointing devices such as a touch panel, track pad, or mouse; a keyboard; imaging means such as a camera or scanner; card readers such as an IC card reader or a magnetic card reader; and audio input means such as a microphone. Examples of the output device 206 include display devices such as an LED display or a liquid crystal display; audio output devices such as a speaker; a printer; etc. In the present disclosure, the input device 205 and the output device 206 are configured separately, but the input device 205 and the output device 206 may also be configured as an integrated device, such as a touch panel display.

[0053] First, an example of the processing of the trained model generation program of the present disclosure will be specifically described with reference to Fig. 6. Fig. 6 is a flowchart showing an example of each procedure of the trained model generation program of the present disclosure.

[0054] First, the learning data acquisition unit 21 acquires the time-series information indicating the characteristics of the target as learning data (S21, learning data acquisition step).

[0055] Next, the trained model generation unit 22 generates a trained model for outputting future risks to the target from the time-series information through machine learning using the training data (S22, trained model generation procedure). The trained model is the trained model of the present disclosure. The trained model of the present disclosure is machine-learned using features, and the features are features that represent time-series trends of the target based on any function converted from the time-series information of the target, and is a trained model for causing a computer to function to output future risks to the target from the time-series information.

[0056] According to the trained model generation program of the present disclosure, the training data acquisition unit acquires the time-series information indicating the characteristics of the target as training data, and the trained model generation unit generates a trained model by machine learning using the training data. This makes it possible to generate a trained model for outputting a risk that may occur to the target in the future from the time-series information.

[0057] Next, the trained model generation method of the present disclosure will be described.

[0058] The trained model generation method of the present disclosure can incorporate the descriptions of the trained model generation program and trained model generation device of the present disclosure. The trained model generation method of the present disclosure is, for example, a method implemented by replacing each "procedure" in the program of the present disclosure with a "step." Specifically, the trained model generation method of the present disclosure includes a training data acquisition step and a trained model generation step. The trained model generation method of the present disclosure can be implemented using, for example, the device 20 of FIG. 4. Note that the trained model generation method of the present disclosure is not limited to use of the device 20 of FIG. 4.

[0059] Third Embodiment Next, other examples of the trained model generation program, trained model generation device, trained model generation method, and trained model of the present disclosure will be described.

[0060] The trained model generation program of the present disclosure is a program that causes a computer to further execute a training data acquisition procedure, a trained model generation procedure, and a data processing procedure. The trained model generation program of the present disclosure can also be said to be a program that causes a computer to function as the training data acquisition procedure, the trained model generation procedure, and the data processing procedure. Furthermore, the trained model generation program of the present disclosure can also be said to be a program that causes a computer to execute, for example, each step of the trained model generation method described below.

[0061] Next, an example of a trained model generation device according to the present disclosure will be described with reference to FIG. 7 .

[0062] 7 is a block diagram showing an example configuration of a trained model generation device 20A (the present device 20A) according to the present disclosure. As shown in FIG. 7, the present device 20A includes a training data acquisition unit 21, a trained model generation unit 22, and a data processing unit 23. Note that the present device 20A is similar to the present device 20 of embodiment 2 except that, in its hardware configuration, a central processing unit 201 functions as the training data acquisition unit 21, the trained model generation unit 22, and the data processing unit 23, and the description thereof can be cited.

[0063] Next, an example of the processing of the trained model generation program of the present disclosure will be specifically described with reference to Fig. 8. Fig. 8 is a flowchart showing an example of each procedure of the trained model generation program of the present disclosure.

[0064] First, the device 20A performs the process of S21 in the same manner as in the second embodiment.

[0065] Next, the data processing unit 23 converts the time series information into any of the functions and performs processing to extract the feature amount that represents the time series trend of the target based on the function (S23, data processing procedure).

[0066] Next, the trained model generation unit 22 generates a trained model using the feature amounts as training data. Other processes are the same as those in S22 of the second embodiment.

[0067] According to the trained model generation program of the present disclosure, the training data acquisition unit acquires the time-series information indicating the characteristics of the target as training data, the data processing unit converts the time-series information into an arbitrary function and performs processing to extract the feature amount indicating the time-series trend of the target based on the function, and the trained model generation unit generates a trained model using the feature amount as training data. This makes it possible to generate a trained model for outputting future risks to the target from the time-series information.

[0068] Next, another example of the trained model generation method of the present disclosure will be described.

[0069] The trained model generation method of the present disclosure can incorporate the descriptions of the trained model generation program and trained model generation device of the present disclosure. The trained model generation method of the present disclosure is, for example, a method implemented by replacing each "procedure" in the program of the present disclosure with a "process." Specifically, the trained model generation method of the present disclosure includes a training data acquisition process, a trained model generation process, and a data processing process. The trained model generation method of the present disclosure can be implemented using, for example, the device 20A of FIG. 7. Note that the trained model generation method of the present disclosure is not limited to use with the device 20A of FIG. 7.

[0070] [Embodiment 4] An example of risk estimation using the risk estimation support device of the present disclosure will be described with specific examples using Figures 9 and 10. Note that, although the present disclosure shows an example using the device 10, the present disclosure is not limited to this. Furthermore, in the present embodiment, an example of estimating a disease onset risk in a human subject has been shown, but the present disclosure is not limited to this.

[0071] First, the apparatus 10 acquires the time-series information, which is, for example, time-series information based on a given time or age for a given test value of a target individual.

[0072] Next, the device 10 calculates a possible risk to the target based on the time-series information. FIG. 9 is a diagram illustrating an example of converting time-series information into a function and a feature space. As shown in FIG. 9, for example, the time-series information is converted into an arbitrary function. In the present disclosure, the function is converted into a linear function. As shown in FIG. 9, the converted linear function includes a slope α (feature 1) and an intercept β (feature 2) as parameters indicating the characteristics of the linear function. These parameters are used as feature quantities, and the data is converted into a feature space representing a time-series trend, for example, with each feature quantity on the vertical axis or horizontal axis. In the case of the present disclosure, the risk is calculated based on, for example, the parameters and the feature space. Note that the risk may be calculated, for example, using a trained model.

[0073] Finally, the device 10 outputs the calculated risk.

[0074] Fifth Embodiment Next, an example of a procedure for generating a trained model according to the present disclosure will be described with a specific example using Fig. 10. Note that, although the present disclosure illustrates an example in which the present device 20A is used, the present disclosure is not limited to this.

[0075] First, the device 20A acquires the time-series information as learning data.

[0076] Next, the device 20A converts the time-series information into an arbitrary function and extracts features representing the time-series trends of the target based on the function, which is the same process as that described in the fourth embodiment, for example.

[0077] Next, the device 20A generates a trained model for outputting future risks to a subject from the time-series information through machine learning using the training data. FIG. 10 is a diagram showing an example of a procedure for generating a trained model according to the present disclosure. Examples of information about the subject include, but are not limited to, feature values ​​in the feature space, whether or not the subject has developed a disease, the period of onset, and covariates such as test values, age, and gender. Next, as shown in FIG. 10, the trained model is generated by performing a survival analysis using the feature values ​​as covariates. Note that the procedure for generating the trained model is merely an example and is not limited thereto.

[0078] [Embodiment 6] The program of the present disclosure may be recorded on, for example, a computer-readable recording medium. The recording medium may be, for example, a non-transitory computer-readable storage medium. The recording medium is not particularly limited, and examples thereof include random access memory (RAM), read-only memory (ROM), hard disk (HD), flash memory (e.g., USB flash memory, SD / SDHC card, etc.), optical disk (e.g., CD-R / CD-RW, DVD-R / DVD-RW, BD-R / BD-RE, etc.), magneto-optical disk (MO), floppy disk (FD), etc. Furthermore, the program of the present disclosure (also referred to as, for example, a programming product or program product) may be distributed from, for example, an external computer. The "distribution" may be, for example, distribution via a communication network or via a device connected via a wire. The program of the present disclosure may be installed and executed on the device to which it is distributed, or may be executed without being installed.

[0079] Although the present disclosure has been described above with reference to the embodiments, the present disclosure is not limited to the above embodiments. The configuration and conditions of the present disclosure may be modified in various ways that are understandable to those skilled in the art within the scope of the present disclosure.

[0080] <Supplementary Notes> Some or all of the above embodiments can be described as in the following supplementary notes, but are not limited to the following. (Supplementary Note 1) A risk estimation support program for causing a computer to execute each of the steps, including an information acquisition step, a risk calculation step, and a risk output step, wherein the information acquisition step acquires time-series information indicating characteristics of a subject, the risk calculation step calculates a future risk that may occur to the subject based on the time-series information, and the risk output step outputs the risk. (Supplementary Note 2) The program according to Supplementary Note 1, wherein the time-series information is time-series information consisting of an arbitrary number of time points. (Supplementary Note 3) The program according to Supplementary Note 1, wherein the time-series information is time-series information for an arbitrary time interval. (Supplementary Note 4) The program according to any of Supplements 1 to 3, wherein the risk is a disease development risk. (Supplementary Note 5) The program according to Supplementary Note 4, wherein the subject is a human, and the time-series information includes time-series information related to health data. (Supplementary Note 6) The program according to any one of Supplements 1 to 5, wherein the risk calculation step converts the time-series information into an arbitrary function and calculates the risk based on the function. (Supplementary Note 7) The program according to Supplementary Note 6, wherein the risk calculation step extracts a feature amount that represents a time-series trend of the target based on the function and calculates the risk based on the feature amount. (Supplementary Note 8) The program according to any one of Supplements 1 to 7, wherein the risk calculation step calculates the risk using a trained model, wherein the trained model is a trained model obtained by machine learning using a feature amount, and the feature amount is a feature amount that represents a time-series trend of the target based on an arbitrary function converted from time-series information of the target. (Supplementary Note 9) A trained model generation program for causing a computer to execute each of the steps, including a training data acquisition step and a trained model generation step, wherein the training data acquisition step acquires time-series information that represents features of the target as training data, and the trained model generation step generates a trained model for outputting a risk that may occur to the target in the future from the time-series information by machine learning using the training data. (Supplementary Note 10) The program according to Supplementary Note 9, wherein the time series information is time series information consisting of an arbitrary number of time points.(Supplementary Note 11) The program according to Supplementary Note 9, wherein the time-series information is time-series information for an arbitrary time interval. (Supplementary Note 12) The program according to any one of Supplements 9 to 11, wherein the risk is a disease onset risk. (Supplementary Note 13) The program according to any one of Supplementary Note 12, wherein the subject is a human, and the time-series information includes time-series information related to health data. (Supplementary Note 14) The program according to any one of Supplementary Notes 9 to 13, comprising a data processing step, wherein the data processing step converts the time-series information into an arbitrary function and performs processing to extract features that represent a time-series trend of the subject based on the function, and wherein the trained model generation step generates a trained model using the features as training data. (Supplementary Note 15) A trained model that is machine-learned using features, wherein the features represent a time-series trend of the subject based on an arbitrary function converted from the time-series information of the subject, and causes a computer to function to output a risk that may occur in the future to the subject from the time-series information. (Supplementary Note 16) A risk estimation support device comprising an information acquisition unit, a risk calculation unit, and a risk output unit, wherein the information acquisition unit acquires time-series information indicating characteristics of a subject, the risk calculation unit calculates a future risk that may occur to the subject based on the time-series information, and the risk output unit outputs the risk. (Supplementary Note 17) The risk estimation support device according to Supplementary Note 16, wherein the time-series information is time-series information consisting of an arbitrary number of time points. (Supplementary Note 18) The risk estimation support device according to Supplementary Note 16, wherein the time-series information is time-series information for an arbitrary time interval. (Supplementary Note 19) The risk estimation support device according to any of Supplements 16 to 18, wherein the risk is a risk of developing a disease. (Supplementary Note 20) The risk estimation support device according to Supplementary Note 19, wherein the subject is a human, and the time-series information includes time-series information related to health data. (Supplementary Note 21) The risk estimation support device according to any one of Supplementary Notes 16 to 20, wherein the risk calculation unit converts the time-series information into an arbitrary function and calculates the risk based on the function. (Supplementary Note 22) The risk estimation support device according to Supplementary Note 21, wherein the risk calculation unit extracts a feature value that represents a time-series trend of the target based on the function and calculates the risk based on the feature value.(Supplementary Note 23) The risk estimation support device according to any one of Supplementary Notes 16 to 22, wherein the risk calculation unit calculates the risk using a trained model, and the trained model is a trained model obtained by machine learning using features, and the features are features that represent a time series trend of the object based on an arbitrary function converted from time series information of the object. (Supplementary Note 24) A trained model generation device including a training data acquisition unit and a trained model generation unit, wherein the training data acquisition unit acquires time series information indicating features of the object as training data, and the trained model generation unit generates a trained model for outputting a risk that may occur to the object in the future from the time series information by machine learning using the training data. (Supplementary Note 25) The trained model generation device according to Supplementary Note 24, wherein the time series information is time series information consisting of an arbitrary number of time points. (Supplementary Note 26) The trained model generation device according to Supplementary Note 24, wherein the time series information is time series information of an arbitrary time interval. (Supplementary Note 27) The trained model generation device according to any one of Supplements 24 to 26, wherein the risk is a risk of developing a disease. (Supplementary Note 28) The trained model generation device according to Supplementary Note 27, wherein the subject is a human, and the time-series information includes time-series information related to health data. (Supplementary Note 29) The trained model generation device according to any one of Supplements 24 to 28, comprising: a data processing unit, wherein the data processing unit converts the time-series information into an arbitrary function and performs processing to extract features that represent a time-series trend of the subject based on the function, and the trained model generation unit generates a trained model using the features as training data. (Supplementary Note 30) A risk estimation support method, comprising an information acquisition step, a risk calculation step, and a risk output step, wherein the information acquisition step acquires time-series information that represents features of the subject, the risk calculation step calculates a future risk that may occur to the subject based on the time-series information, and the risk output step outputs the risk, wherein each of the steps is executed by a computer. (Supplementary Note 31) The risk estimation support method according to Supplementary Note 30, wherein the time series information is time series information consisting of an arbitrary number of time points. (Supplementary Note 32) The risk estimation support method according to Supplementary Note 30, wherein the time series information is time series information at an arbitrary time interval.(Supplementary Note 33) The risk estimation support method according to any one of Supplements 30 to 32, wherein the risk is a disease onset risk. (Supplementary Note 34) The risk estimation support method according to Supplementary Note 33, wherein the subject is a human, and the time-series information includes time-series information related to health data. (Supplementary Note 35) The risk estimation support method according to any one of Supplements 30 to 34, wherein the risk calculation step converts the time-series information into an arbitrary function and calculates the risk based on the function. (Supplementary Note 36) The risk estimation support method according to Supplementary Note 35, wherein the risk calculation step extracts a feature amount representing a time-series trend of the subject based on the function, and calculates the risk based on the feature amount. (Supplementary Note 37) The risk estimation support method according to any one of Supplements 30 to 36, wherein the risk calculation step calculates the risk using a trained model, wherein the trained model is a trained model trained by machine learning using a feature amount, and the feature amount is a feature amount representing a time-series trend of the subject based on an arbitrary function converted from the time-series information of the subject. (Supplementary Note 38) A trained model generation method comprising a training data acquisition step and a trained model generation step, wherein the training data acquisition step acquires time series information indicating characteristics of a subject as training data, and the trained model generation step generates a trained model for outputting a future risk that may occur to the subject from the time series information by machine learning using the training data, wherein each of the steps is executed by a computer. (Supplementary Note 39) The trained model generation method according to Supplementary Note 38, wherein the time series information is time series information consisting of an arbitrary number of time points. (Supplementary Note 40) The trained model generation method according to Supplementary Note 38, wherein the time series information is time series information for an arbitrary time interval. (Supplementary Note 41) The trained model generation method according to any of Supplements 38 to 40, wherein the risk is a risk of developing a disease. (Supplementary Note 42) The trained model generation method according to Supplementary Note 41, wherein the subject is a human, and the time series information includes time series information related to health data.(Supplementary Note 43) A trained model generation method according to any one of Supplements 38 to 42, including a data processing step, wherein the data processing step converts the time series information into an arbitrary function and performs processing to extract features that represent time series trends of the object based on the function, and a trained model generation step generates a trained model using the features as training data. (Supplementary Note 44) A computer-readable recording medium having recorded thereon a risk estimation support program for causing a computer to execute each of the steps, including an information acquisition step, a risk calculation step, and a risk output step, wherein the information acquisition step acquires time series information that represents features of the object, the risk calculation step calculates a risk that may occur in the future to the object based on the time series information, and the risk output step outputs the risk. (Supplementary Note 45) The recording medium according to Supplementary Note 44, wherein the time series information is time series information consisting of an arbitrary number of time points. (Supplementary Note 46) The recording medium according to Supplementary Note 44, wherein the time series information is time series information for an arbitrary time interval. (Appendix 47) The recording medium of any of Appendices 44 to 46, wherein the risk is a disease development risk. (Appendix 48) The recording medium of Appendices 47, wherein the subject is a human, and the time-series information includes time-series information related to health data. (Appendix 49) The recording medium of any of Appendices 44 to 48, wherein the risk calculation step converts the time-series information into an arbitrary function and calculates the risk based on the function. (Appendix 50) The recording medium of Appendices 49, wherein the risk calculation step extracts features representing a time-series trend of the subject based on the function, and calculates the risk based on the features. (Appendix 51) The recording medium of any of Appendices 44 to 50, wherein the risk calculation step calculates the risk using a trained model, wherein the trained model is a trained model trained by machine learning using features, and the features are features representing a time-series trend of the subject based on an arbitrary function converted from the time-series information of the subject.(Appendix 52) A computer-readable recording medium having recorded thereon a trained model generation program for causing a computer to execute each of the steps, the trained data acquisition step acquiring time-series information indicating characteristics of a subject as training data, and the trained model generation step generating a trained model from the time-series information by machine learning using the training data, for outputting a future risk that may occur to the subject. (Appendix 53) The recording medium of Appendix 52, wherein the time-series information is time-series information consisting of an arbitrary number of time points. (Appendix 54) The recording medium of Appendix 52, wherein the time-series information is time-series information for an arbitrary time interval. (Appendix 55) The recording medium of any of Appendices 52 to 54, wherein the risk is a disease development risk. (Appendix 56) The recording medium of Appendix 55, wherein the subject is a human, and the time-series information includes time-series information related to health data. (Appendix 57) A recording medium described in any of Appendices 52 to 56, including a data processing procedure that converts the time series information into an arbitrary function and performs processing to extract features that represent trends in the time series of the target based on the function, and a trained model generation procedure that generates a trained model using the features as training data.

[0081] This application claims priority based on Japanese Patent Application No. 2024-102909, filed on June 26, 2024, the disclosure of which is incorporated herein in its entirety by reference.

[0082] According to the present disclosure, it is possible to estimate future risks to a subject based on the subject's time-series information. The fields to which the present disclosure can be applied are not limited, and the program of the present disclosure is useful in various fields.

[0083] REFERENCE SIGNS LIST 10 Risk estimation support device 11 Information acquisition unit 12 Risk calculation unit 13 Risk output unit 20, 20A Trained model generation device 21 Training data acquisition unit 22 Trained model generation unit 23 Data processing unit 101, 201 CPU 102, 202 Memory 103, 203 Bus 104, 204 Storage device 105, 205 Input device 106, 206 Output device 107, 207 Communication device

Claims

1. A risk estimation support program for causing a computer to execute each of the above procedures, including an information acquisition procedure, a risk calculation procedure, and a risk output procedure, wherein the information acquisition procedure acquires time-series information indicating characteristics of an object, the risk calculation procedure calculates future risks that may occur to the object based on the time-series information, and the risk output procedure outputs the risks.

2. The program according to claim 1, wherein the time series information is time series information consisting of an arbitrary number of time points.

3. The program according to claim 1, wherein the time series information is time series information for an arbitrary time interval.

4. The program according to claim 1, wherein the risk is a risk of developing a disease.

5. The program according to claim 4, wherein the subject is a human, and the time-series information includes time-series information related to health data.

6. The program according to any one of claims 1 to 5, wherein the risk calculation step converts the time-series information into an arbitrary function and calculates the risk based on the function.

7. The program according to claim 6, wherein the risk calculation step extracts a feature value that represents a time-series trend of the target based on the function, and calculates the risk based on the feature value.

8. A program according to any one of claims 1 to 5, wherein the risk calculation procedure calculates the risk using a trained model, the trained model being a trained model machine-learned using features, and the features being features that represent the time series trend of the target based on an arbitrary function converted from the time series information of the target.

9. A risk estimation support device comprising an information acquisition unit, a risk calculation unit, and a risk output unit, wherein the information acquisition unit acquires time-series information indicating characteristics of an object, the risk calculation unit calculates future risks that may occur to the object based on the time-series information, and the risk output unit outputs the risks.

10. A risk estimation support method comprising an information acquisition step, a risk calculation step, and a risk output step, wherein the information acquisition step acquires time-series information indicating characteristics of an object, the risk calculation step calculates future risks that may occur to the object based on the time-series information, and the risk output step outputs the risks, and wherein each of the steps is executed by a computer.

11. A computer-readable recording medium having recorded thereon a risk estimation support program for causing a computer to execute each of the above procedures, including an information acquisition procedure, a risk calculation procedure, and a risk output procedure, wherein the information acquisition procedure acquires time-series information indicating characteristics of an object, the risk calculation procedure calculates future risks that may occur to the object based on the time-series information, and the risk output procedure outputs the risks.

12. A trained model generation method comprising: a training data acquisition step; and a trained model generation step, wherein the training data acquisition step acquires time-series information indicating characteristics of a target as training data; and the trained model generation step generates a trained model for outputting future risks to the target from the time-series information by machine learning using the training data, wherein each of the steps is executed by a computer.

13. The trained model generation method according to claim 12, wherein the time series information is time series information consisting of an arbitrary number of time points.

14. The trained model generation method according to claim 12, wherein the time series information is time series information at any time interval.

15. A trained model generation method described in any one of claims 12 to 14, wherein the risk is a risk of developing a disease.

16. The trained model generation method according to claim 15, wherein the subject is a human, and the time series information includes time series information related to health data.

17. A trained model generation method according to any one of claims 12 to 14, further comprising a data processing step of converting the time series information into an arbitrary function and extracting features representing trends in the time series of the target based on the function, and a trained model generation step of generating a trained model using the features as training data.

18. A trained model generation device including a training data acquisition unit and a trained model generation unit, wherein the training data acquisition unit acquires time series information indicating characteristics of a target as training data, and the trained model generation unit generates a trained model for outputting future risks to the target from the time series information through machine learning using the training data.

19. A trained model generation program for causing a computer to execute each of the above steps, including a training data acquisition step for acquiring time-series information indicating characteristics of a target as training data, and a trained model generation step for generating a trained model from the time-series information by machine learning using the training data, for outputting future risks that may occur to the target.

20. A computer-readable recording medium having recorded thereon a trained model generation program for causing a computer to execute each of the above steps, the trained data acquisition step acquiring time-series information indicating characteristics of a target as training data, and the trained model generation step generating a trained model from the time-series information by machine learning using the training data to output risks that may occur to the target in the future.

Images