JP7692815B2 - Visualization device, visualization method, and visualization program - Google Patents

Description

The present invention relates to a visualization device, a visualization method, and a visualization program.

In recent years, there has been a growing expectation for technology that utilizes machine learning models that have learned the characteristics of time-series data from sensors and other sources generated in various locations in urban spaces, including plants. Conventionally, the machine learning models that are actually applied have been selected based on statistical indicators such as prediction accuracy.

JP 2019-185194 A

However, a machine learning model selected based on statistical indicators is not necessarily the most suitable machine learning model in the field. For example, in the field, there is a tendency to be convinced that a machine learning model is the most suitable model by checking the changes in the explanatory variables that are the cause of changes in the target variable of the machine learning model based on knowledge and know-how.

The present invention has been made in consideration of the above, and aims to make it possible to check changes in explanatory variables that are factors in changes in the objective variable of a machine learning model, and to select the optimal machine learning model based on the user's knowledge and know-how.

In order to solve the above-mentioned problems and achieve the object, the visualization device according to the present invention is characterized by having an acquisition unit that acquires time-series data, a generation unit that learns using the acquired time-series data to generate a plurality of models, and a presentation unit that presents the causal relationship between each explanatory variable and a target variable in a time series for each generated model.

According to the present invention, it is possible to check the changes in explanatory variables that are the cause of changes in the objective variable of a machine learning model, and to select the optimal machine learning model based on the user's knowledge and know-how.

FIG. 1 is a schematic diagram illustrating a schematic configuration of a visualization device. FIG. 2 is a diagram for explaining the process of the presentation unit. FIG. 3 is a diagram for explaining the process of the presentation unit. FIG. 4 is a flowchart illustrating a visualization process procedure. FIG. 5 is a diagram illustrating a computer that executes a visualization program.

One embodiment of the present invention will be described in detail below with reference to the drawings. Note that the present invention is not limited to this embodiment. In addition, in the drawings, the same parts are denoted by the same reference numerals.

[Configuration of visualization device]
Fig. 1 is a schematic diagram illustrating a schematic configuration of a visualization device. As illustrated in Fig. 1, a visualization device 10 is realized by a general-purpose computer such as a personal computer, and includes an input unit 11, an output unit 12, a communication control unit 13, a storage unit 14, and a control unit 15.

The input unit 11 is realized using input devices such as a keyboard and a mouse, and inputs various instruction information, such as a command to start processing, to the control unit 15 in response to an input operation by an operator. The output unit 12 is realized by a display device such as a liquid crystal display, a printing device such as a printer, etc.

The communication control unit 13 is realized by a NIC (Network Interface Card) or the like, and controls communication between the control unit 15 and an external device via a network. For example, the communication control unit 13 controls communication between the control unit 15 and a sensor that outputs time series data to be processed in the visualization process described below, a management device that manages the time series data, or the like.

The storage unit 14 is realized by a semiconductor memory element such as a random access memory (RAM) or a flash memory, or a storage device such as a hard disk or an optical disk. The storage unit 14 stores in advance the processing program that operates the visualization device 10 and data used during the execution of the processing program, or stores the data temporarily each time processing is performed.

In this embodiment, the storage unit 14 stores, for example, models and model parameters generated in the visualization process described below. The storage unit 14 may be configured to communicate with the control unit 15 via the communication control unit 13.

The control unit 15 is realized using a CPU (Central Processing Unit) or the like, and executes a processing program stored in memory. As a result, the control unit 15 functions as an acquisition unit 15a, a generation unit 15b, a presentation unit 15c, a calculation unit 15d, and an identification unit 15e, as exemplified in FIG. 1. Note that each or some of these functional units may be implemented in different hardware. For example, the generation unit 15b may be implemented in a device separate from the other functional units. The control unit 15 may also include other functional units.

The acquisition unit 15a acquires time series data. For example, the acquisition unit 15a acquires time series data such as sensor values to be subjected to the visualization process described below via the input unit 11, or from a sensor, or from a management device that manages the sensor values, via the communication control unit 13.

The generation unit 15b learns using the acquired time series data and generates multiple models with different parameter values. Specifically, the generation unit 15b uses the acquired time series data as teacher data to generate multiple models that predict a target variable from multiple explanatory variables by changing the parameter values.

The presentation unit 15c presents the causal relationship between each explanatory variable and the objective variable in a time series for each generated model. Here, FIG. 2 and FIG. 3 are diagrams for explaining the processing of the presentation unit 15c. For example, as illustrated in FIG. 2, the presentation unit 15c presents to the user the time series change in the causal relationship between the value of each explanatory variable of each model and the value of the objective variable predicted using the model. Note that the causal relationship includes correlations, attributions that represent causal relationships of actions, and the like.

Figure 2 illustrates the time series changes in the factor relationships between the explanatory variable values of features A, B, ..., I and the objective variable for each model, Model 1, Model 2, ..., Model N. In the example shown in Figure 2, the greater the factor relationship, the darker the color displayed. The presentation unit 15c outputs the factor relationships between each explanatory variable and the objective variable of each model to a user terminal or the like via the output unit 12 or the communication control unit 13, and presents them to the user.

By checking the presented results, users can confirm the explanatory variables that are the cause of changes in the target variable in each model. This allows users to confirm and select models that are in line with their own knowledge and know-how.

The presenting unit 15c may also accept input of an evaluation value by the user for each model. For example, the user specifies and inputs an evaluation value so that the closer a model is to the user's knowledge and know-how, the higher the evaluation value. In this way, a high evaluation value is set according to the satisfaction of the user who has knowledge and know-how.

The presentation unit 15c may further present the causal relationships between the explanatory variables for each model, as illustrated in FIG. 3. In the example shown in FIG. 3, in addition to the causal relationships between the explanatory variables and the objective variable of each model, a causal graph showing the causal relationships between the explanatory variables and the time series changes in the prediction results of the model are presented. This allows the user to check the causal relationships between the explanatory variables and the objective variable in more detail, making it easier for the user to confirm and select an optimal model based on his or her own knowledge and know-how.

Returning to the explanation of FIG. 1, the calculation unit 15d calculates an evaluation value that represents the smoothness of change in the time series direction of the factor relationship between each of the presented explanatory variables and the objective variable.

The smoother the change in the factorial relationship between the explanatory variable and the objective variable over time, the easier it is to perceive it as a factor in the change in the objective variable, and the more convinced knowledgeable users will be. Therefore, by evaluating explanatory variables whose factorial relationship with the objective variable over time changes smoothly as major factors in the change in the objective variable, it is possible to highly evaluate explanatory variables that are highly convinced by users.

The calculation unit 15d therefore calculates the smoothness of change in the time series direction of the factor relationship between each explanatory variable and the objective variable as an evaluation value. For example, the calculation unit 15d calculates the evaluation value so that the smaller the change in the factor relationship between adjacent times in the time series data, the higher the evaluation value representing the smoothness of change. In the example shown in FIG. 2, the smoother the color change, the higher the evaluation value.

This makes it possible to evaluate explanatory variables whose factor relationships with the objective variable change smoothly over time as major factors in the change of the objective variable. This allows users to objectively evaluate models that are highly convincing based on their knowledge and know-how.

The identification unit 15e selects a model based on the evaluation value. Specifically, the identification unit 15e selects, for example, the model with the maximum evaluation value as the optimal model based on the evaluation value input by the user or the calculated evaluation value. The identification unit 15e then outputs the selected model to a user terminal or the like via the output unit 12 or the communication control unit 13.

In a real environment, it is not necessary to specify one model, and multiple models may be applied and run side by side. Alternatively, models may be switched and applied depending on the situation.

[Visualization Processing Procedure]
Next, an example of visualization processing by the visualization device 10 according to the present embodiment will be described with reference to Fig. 4. Fig. 4 is a flowchart illustrating a visualization processing procedure. The flowchart in Fig. 4 is started, for example, when an input is received to instruct the start of the visualization processing.

First, the acquisition unit 15a acquires time series data. For example, the acquisition unit 15a acquires time series data such as sensor values to be subjected to visualization processing (step S1).

Next, the generation unit 15b learns using the acquired time series data and generates multiple models with different parameter values (step S2). Specifically, the generation unit 15b uses the acquired time series data as teacher data to generate multiple models that predict a target variable from multiple explanatory variables by changing the parameter values.

Next, the presentation unit 15c presents the causal relationship between each explanatory variable and the objective variable in a time series for each generated model (step S3). For example, the presentation unit 15c presents to the user the time series change in the causal relationship between the value of each explanatory variable of each model and the value of the objective variable predicted using the model.

Then, the visualization device 10 evaluates each model (step S4). For example, the presentation unit 15c accepts an input of an evaluation value by the user. Alternatively, the calculation unit 15d calculates an evaluation value that represents the smoothness of the change.

The identification unit 15e also selects a model based on the evaluation value. For example, the model with the highest evaluation value is selected as the optimal model. This completes the series of visualization processes.

[effect]
As described above, in the visualization device 10 of the above embodiment, the acquisition unit 15a acquires time-series data. The generation unit 15b learns using the acquired time-series data and generates a plurality of models with different parameter values. The presentation unit 15c presents, for each generated model, a causal relationship between each explanatory variable and a target variable in a time series.

This allows users at a plant or other site, for example, to check the explanatory variables that are the cause of changes in the target variable in each model. This allows users to check and select the optimal model based on their own knowledge and know-how. It also makes it possible to present models to users and have them select in a shorter period of time than it would take for a developer to select the optimal model and present it to the user.

The presentation unit 15c also accepts user input of evaluation values for each model. This makes it possible to evaluate each model while reflecting the user's knowledge and know-how.

In addition, the presentation unit 15c further presents the causal relationships between each explanatory variable for each model. This makes it easier for the user to confirm the causal relationships between each explanatory variable and the objective variable in more detail and to confirm or select the optimal model based on the user's own knowledge and know-how.

The calculation unit 15d also calculates an evaluation value that represents the smoothness of change in the time series direction of the factor relationship between each of the presented explanatory variables and the objective variable. This makes it possible to evaluate, for example, an explanatory variable whose factor relationship with the objective variable changes smoothly in the time series direction as a major factor in the change in the objective variable. Therefore, it becomes possible to objectively evaluate a model that is convincing to the user based on knowledge and know-how as an optimal model.

[System configuration, etc.]
Each component of each device shown in the figure is a functional concept, and does not necessarily have to be physically configured as shown in the figure. In other words, the specific form of distribution and integration of each device is not limited to that shown in the figure, and all or a part of it can be functionally or physically distributed and integrated in any unit depending on various loads and usage conditions. Furthermore, each processing function performed by each device can be realized in whole or in part by a CPU or GPU and a program analyzed and executed by the CPU or GPU, or can be realized as hardware using wired logic.

Furthermore, among the processes described in this embodiment, all or part of the processes described as being performed automatically can be performed manually, or all or part of the processes described as being performed manually can be performed automatically using known methods. In addition, the information including the processing procedures, control procedures, specific names, various data, and parameters shown in the above documents and drawings can be changed as desired unless otherwise specified.

[program]
It is also possible to create a program in which the processing performed by the visualization device described in the above embodiment is written in a language executable by a computer. For example, it is also possible to create a program in which the processing performed by the visualization device 10 according to the embodiment is written in a language executable by a computer. In this case, the same effect as in the above embodiment can be obtained by having the computer execute the program. Furthermore, such a program may be recorded on a computer-readable recording medium, and the program recorded on the recording medium may be read and executed by a computer to realize the same processing as in the above embodiment.

Figure 5 is a diagram showing an example of a computer that executes a visualization program. The computer 1000 has, for example, a memory 1010, a CPU 1020, a hard disk drive interface 1030, a disk drive interface 1040, a serial port interface 1050, a video adapter 1060, and a network interface 1070. These components are connected by a bus 1080.

The memory 1010 includes a ROM (Read Only Memory) 1011 and a RAM 1012. The ROM 1011 stores a boot program such as a BIOS (Basic Input Output System). The hard disk drive interface 1030 is connected to a hard disk drive 1031. The disk drive interface 1040 is connected to a disk drive 1041. A removable storage medium such as a magnetic disk or optical disk is inserted into the disk drive 1041. The serial port interface 1050 is connected to a mouse 1051 and a keyboard 1052, for example. The video adapter 1060 is connected to a display 1061, for example.

Here, the hard disk drive 1031 stores, for example, an OS 1091, an application program 1092, a program module 1093, and program data 1094. Each piece of information described in the above embodiment is stored, for example, in the hard disk drive 1031 or memory 1010.

The visualization program is stored in the hard disk drive 1031, for example, as a program module 1093 in which instructions to be executed by the computer 1000 are written. Specifically, the program module 1093 in which each process executed by the visualization device 10 described in the above embodiment is written is stored in the hard disk drive 1031.

In addition, data used for information processing by the visualization program is stored as program data 1094, for example, in the hard disk drive 1031. Then, the CPU 1020 reads the program module 1093 and program data 1094 stored in the hard disk drive 1031 into the RAM 1012 as necessary, and executes each of the procedures described above.

The program module 1093 and program data 1094 related to the visualization program are not limited to being stored in the hard disk drive 1031, and may be stored in a removable storage medium, for example, and read by the CPU 1020 via the disk drive 1041 or the like. Alternatively, the program module 1093 and program data 1094 related to the visualization program may be stored in another computer connected via a network, such as a LAN (Local Area Network) or a WAN (Wide Area Network), and read by the CPU 1020 via the network interface 1070.

The above describes an embodiment of the invention made by the inventor, but the present invention is not limited to the description and drawings that form part of the disclosure of the present invention according to this embodiment. In other words, other embodiments, examples, operational techniques, etc. made by those skilled in the art based on this embodiment are all included in the scope of the present invention.

REFERENCE SIGNS LIST 10 Visualization device 11 Input unit 12 Output unit 13 Communication control unit 14 Storage unit 15 Control unit 15a Acquisition unit 15b Generation unit 15c Presentation unit 15d Calculation unit 15e Identification unit

Claims (5)

An acquisition unit that acquires time series data;
A generation unit that learns using the acquired time-series data and generates a plurality of models with different parameter values;
a presentation unit that presents, for each generated model, a factor relationship between each explanatory variable and a target variable in a time series;
a calculation unit that calculates an evaluation value that represents the smoothness of change in the time series direction of the factor relationship between each of the presented explanatory variables and the objective variable;
A visualization device comprising: The visualization device according to claim 1, characterized in that the presentation unit accepts input of an evaluation value for each model by a user. The visualization device according to claim 1, characterized in that the presentation unit further presents causal relationships between each explanatory variable for each model. A visualization method executed by a visualization device, comprising:
An acquisition step of acquiring time series data;
A generation step of generating a plurality of models having different parameter values by learning using the acquired time-series data;
a presentation step of presenting, for each generated model, a causal relationship between each explanatory variable and a target variable in a time series;
A calculation step of calculating an evaluation value representing the smoothness of change in the time series direction of the factor relationship between each of the presented explanatory variables and the objective variable;
A visualization method comprising: An acquisition step of acquiring time series data;
A generation step of generating a plurality of models having different parameter values by learning using the acquired time-series data;
A presentation step of presenting a factor relationship between each explanatory variable and a target variable in a time series for each generated model;
A calculation step of calculating an evaluation value representing the smoothness of change in the time series direction of the factor relationship between each of the presented explanatory variables and the objective variable;
A visualization program for running the above on a computer.

Images