Machine learning system and machine learning method

Abstract

To enable machine learning with simple operations.SOLUTION: A machine learning system comprises a display device and a learning device. The learning device causes a variable selection section (83) in a learning execution screen (70a) of the display device to display a result of extracting explanatory variables candidates and objective variables candidates from a target file designated by a user. The learning device causes a result display section (85) in the learning execution screen (70a) to display a result of analyzing, by machine learning, a relation between an explanatory variable and an objective variable selected by the user from among the explanatory variables candidates and the objective variables candidates displayed in the variable selection section (83).SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The present disclosure relates to a machine learning system and a machine learning method.

Background Art

[0002] For example, Japanese Patent Application Laid-Open No. 2023-34011 (Patent Document 1) describes a technique for optimizing control parameters in a battery manufacturing process using machine learning.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Generally, in machine learning, the relationship between explanatory variables (inputs) and target variables (outputs) is learned using a predetermined algorithm. Conventionally, performing machine learning has required complex coding and specialized knowledge and has not been user-friendly.

[0005] The present disclosure has been made to solve the above-described problems, and an object thereof is to enable machine learning to be performed by a simple operation.

Means for Solving the Problems

[0006] (Item 1) A machine learning system according to the present disclosure includes a display device and a learning device connected to the display device. The learning device causes the display device to display a result of extracting explanatory variable candidates and target variable candidates from a target file specified by a user, and causes the display device to display a result of analyzing, by machine learning, the relationship between the explanatory variables and target variables selected by the user from among the explanatory variable candidates and target variable candidates displayed on the display device.

[0007] (2) In the machine learning system according to claim 1, the learning device generates a plurality of prediction models for predicting the target variable from the explanatory variables by performing machine learning on the explanatory variables and the target variable selected by the user according to a plurality of algorithms, respectively, calculates the coefficient of determination of each of the generated plurality of prediction models, and causes the display device to display the prediction model with the highest coefficient of determination as the best model.

[0008] (3) In the machine learning system according to claim 2, the learning device causes the display device to display a plurality of algorithms, and causes the display device to display the information of the best model or the information of the selected model, which is a prediction model according to the algorithm selected by the user from among the plurality of algorithms displayed on the display device.

[0009] (4) In the machine learning system according to claim 3, the learning device acquires the explanatory variables input by the user, and causes the display device to display the result of calculating the target variable corresponding to the explanatory variables input by the user using the best model or the selected model.

[0010] (5) In the machine learning system according to any one of claims 1 to 4, the learning device acquires the division ratio of the training data and the test data requested by the user, divides the data included in the target file into the training data and the test data according to the acquired division ratio, and causes the display device to display the result of performing machine learning.

[0011] (6) In the machine learning system according to any one of claims 1 to 4, the explanatory variable is a variable related to the material of the battery, and the target variable is a variable related to the characteristics of the battery.

[0012] (Item 7) The machine learning method according to the present disclosure includes a step of causing a display device to display a result of extracting explanatory variable candidates and objective variable candidates from a target file specified by a user, and a step of causing the display device to display a result of analyzing the relationship between the explanatory variables and objective variables selected by the user from among the explanatory variable candidates and objective variable candidates displayed on the display device by machine learning.

Effect of the Invention

[0013] According to the present disclosure, machine learning can be performed by a simple operation without performing a complicated operation such as coding.

Brief Description of the Drawings

[0014]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Mode for Carrying Out the Invention

[0015] Embodiments of the present disclosure will be described in detail with reference to the drawings. In the drawings, the same or corresponding parts are denoted by the same reference numerals and their description will not be repeated.

[0016] [System Configuration] FIG. 1 is a diagram schematically showing the overall configuration of a machine learning system 1 according to the present embodiment.

[0017] The machine learning system 1 includes a display device 2, a keyboard 3, a mouse 4, a removable disk 5, and a learning device 6.

[0018] The learning device 6 is a device that performs machine learning. The learning device 6 includes a support device 10 and an analysis device 60. The analysis device 60 is a device that actually performs machine learning and analyzes the learning results. The support device 10 is a device for supporting the use of machine learning by the analysis device 60.

[0019] As main hardware elements, the support device 10 includes an arithmetic unit 11, a storage unit 12, a display interface 15, a peripheral device interface 16, a reading unit 17, and a communication unit 18. Note that the support device 10 may be realized by, for example, a general-purpose computer or a dedicated computer for machine learning support.

[0020] The arithmetic unit 11 is an arithmetic circuit (arithmetic device) that executes various processes by executing various programs, and is an example of a computer. The arithmetic unit 11 is composed of, for example, a CPU (Central Processing Unit). All or part of the functions of the arithmetic unit 11 may be provided in a server device (for example, a cloud-type server device) not shown in the figure.

[0021] When the arithmetic unit 11 executes an arbitrary program, the storage unit 12 stores program codes, work memories, and the like. The storage unit 12 stores learning data 13 used for the machine learning of the analysis device 60. Note that the learning data 13 includes a plurality of combinations of explanatory variables and target variables.

[0022] The display interface 15 is an interface for connecting the display device 2, and realizes input / output of data between the support device 10 and the display device 2.

[0023] The peripheral device interface 16 is an interface for connecting peripheral devices such as the keyboard 3 and the mouse 4, and realizes input / output of data between the support device 10 and the peripheral devices.

[0024] The reading unit 17 reads various data stored in the removable disk 5 which is a storage medium. The learning data 13 is, for example, read from the removable disk 5 by the reading unit 17 and stored in the storage unit 12.

[0025] The communication unit 18 transmits and receives data to and from the analysis device 60 via wired communication or wireless communication.

[0026] When the analysis device 60 receives a learning command from the support device 10, it performs machine learning based on the received learning command and analyzes the learning results. The learning command includes, in addition to the above-described learning data 13, information on the division ratio of the learning data 13, etc. The division ratio of the learning data 13 is the ratio for dividing a plurality of combined data of the explanatory variable and the objective variable included in the learning data 13 into training data (data used for generating the prediction model 61) and test data (data used for verifying the prediction model 61).

[0027] The analysis device 60 can generate a prediction model 61 using machine learning (learning phase), or perform various processes using the generated prediction model 61 (utilization phase). The prediction model 61 is a model for predicting the objective variable from the explanatory variable. When the explanatory variable is input, the prediction model 61 outputs the value of the objective variable corresponding to the input explanatory variable.

[0028] When the analysis device 60 receives a learning command from the support device 10 during the learning phase, it reads the learning data 13 included in the learning command and automatically performs preprocessing of the learning data. For example, as one of the preprocessing steps, when there are missing values in the read learning data 13, the analysis device 60 performs a process of supplementing the numerical values with similar data. After performing the preprocessing, the analysis device 60 divides the learning data into training data and test data at the division ratio included in the learning command. Then, the analysis device 60 generates a prediction model 61 that predicts the relationship between the explanatory variable and the objective variable included in the training data. Note that, for example, the explanatory variable is a variable related to the material of the battery, and the objective variable is a variable related to the characteristics of the battery. By setting the variables in this way, a prediction model 61 that predicts the relationship between the material of the battery and the characteristics of the battery can be generated.

[0029] The prediction model 61 is generated using a plurality of algorithms respectively. That is, the analysis device 60 generates a plurality of prediction models 61 using a plurality of algorithms. Then, the analysis device 60 uses the test data to verify the generated prediction model 61. Furthermore, the analysis device 60 performs an analysis (such as calculating the correlation coefficient and the coefficient of determination) of the generated prediction model 61. Note that since the generation method, verification method, and analysis method of the prediction model by the analysis device 60 are the same as known methods, detailed descriptions will not be repeated here.

[0030] When the analysis device 60 receives a utilization command including the value of the explanatory variable from the support device 10 during the utilization phase, it calculates the predicted value of the objective variable for the value of the explanatory variable included in the utilization command using the prediction model 61, and transmits the calculated predicted value of the objective variable to the support device 10.

[0031] [Support for Machine Learning] Conventional machine learning tools required specialized knowledge such as complex coding and the process of machine learning. However, it was difficult for general users to acquire specialized knowledge of machine learning in advance, so they were not user-friendly.

[0032] Therefore, the learning device 6 according to the present embodiment supports the user to easily execute machine learning and utilize the machine learning results with simple operations.

[0033] <Learning phase (phase for executing machine learning)> FIG. 2 is a diagram schematically showing an example of a learning execution screen 70a displayed on the display device 2. The learning device 6 causes the display device 2 to display the learning execution screen 70a shown in FIG. 2. The user can perform machine learning by the learning device 6 by simply performing operations according to the display content of the learning execution screen 70a without performing complicated coding or the like. Hereinafter, the display content of the learning execution screen 70a and the operations performed by the user while viewing the learning execution screen 70a will be described.

[0034] (1) Specification of target file At the upper part of the learning execution screen 70a, a tab 80 described as "Search for prediction model" and a tab 90 described as "Display of predicted value" are displayed so as to be selectable. When the user performs an operation of selecting the tab 80, the learning execution screen 70a shown in FIG. 2 is displayed on the screen of the display device 2.

[0035] Below the tabs 80 and 90, a file selection icon 81a, a path display section 81b, and a load icon 81c are displayed. When the user moves the cursor to the file selection icon 81a and clicks it, the storage destination and file name of the learning data (the above-described learning data 13) are automatically searched and displayed in the path display section 81b. When there are multiple files of the learning data, the storage destinations and file names of the multiple files are displayed in the path display section 81b. By moving the cursor to the portion where the intended file name is displayed and clicking it, the user can easily specify the target file to be used for machine learning. When the user moves the cursor to the load icon 81c and clicks it in the state where the target file is specified (hereinafter also referred to as "file load operation"), the loading of the target file is started, and the status bar 81d is popped up and displayed until the loading of the target file is completed. In the status bar 81d, for example, the loading information of the target file, the loading time, etc. are displayed.

[0036] (2) Selection of the learning data splitting method and splitting ratio Below the file selection icon 81a, a split selection section 82 is displayed. The user can select the splitting method and splitting ratio of the learning data included in the loaded target file according to the display of the split selection section 82.

[0037] Specifically, the split selection section 82 includes a method selection section 82a in which options for the splitting method are displayed. In FIG. 2, as options for the splitting method, "Random" for randomly splitting the learning data and "Line number retention" for splitting while maintaining the order of the learning data are displayed, and a state where "Random" is selected by default is shown. When the splitting method may remain the default "Random", no particular operation is required. By moving the cursor to the display portion of "Line number retention" and clicking it, the user can easily change the splitting method to "Line number retention".

[0038] Furthermore, the division selection unit 82 includes a ratio selection unit 82b that displays the division ratio with a scroll bar. In FIG. 2, an example is shown in which the default value of the division ratio (the ratio of training data to the entire learning data) is set to "80%". Note that it is common to set the division ratio between 70% and 90%. The user can arbitrarily change the division ratio by operating the scroll bar of the ratio selection unit 82b.

[0039] (3) Selection of explanatory variables and target variables Below the division selection unit 82, a variable selection unit 83 is displayed. When the loading of the target file is completed, a process of extracting explanatory variable candidates and target variable candidates from the learning data included in the target file is performed, and the title names of the extracted explanatory variable candidates and target variable candidates are respectively displayed selectably in the explanatory variable selection unit 83b and the target variable selection unit 83a.

[0040] The user can easily select an explanatory variable by performing an operation of selecting and clicking one or more intended explanatory variables from the explanatory variable candidates displayed in the explanatory variable selection unit 83b (hereinafter also referred to as "explanatory variable selection operation"). Similarly, the user can easily select a target variable by performing an operation of selecting and clicking one intended target variable from the target variable candidates displayed in the target variable selection unit 83a (hereinafter also referred to as "target variable selection operation"). The explanatory variables and target variables selected by the user are displayed in the selected variable display unit 83c. Hereinafter, the explanatory variable selection operation and the target variable selection operation are also collectively referred to as "variable selection operation".

[0041] Note that the explanatory variable candidates displayed in the explanatory variable selection unit 83b can be added later, and it is also possible to update the machine learning result based on the added explanatory variables.

[0042] (4) Execution of machine learning (generation of prediction model) Below the variable selection section 83, a learning execution icon 84 is displayed. When the user places the cursor on the learning execution icon 84 and clicks it (hereinafter also referred to as a "learning execution operation") while the explanatory variable and the objective variable selected by the user are displayed in the selection variable display section 83c, machine learning by the learning device 6 is executed under the learning conditions (learning data, explanatory variable, objective variable, division method, division ratio, etc.) specified by the user on the learning execution screen 70a.

[0043] The learning device 6 generates a prediction model (the above-mentioned prediction model 61) for predicting the objective variable from the explanatory variables by performing machine learning using the learning conditions specified by the user. As described above, when an explanatory variable is input, the prediction model outputs the value of the objective variable corresponding to the input explanatory variable.

[0044] The learning device 6 according to the present embodiment generates a plurality of prediction models using a plurality of algorithms for the learning conditions specified by the user. Further, the learning device 6 calculates a correlation coefficient (an index representing the relationship between the explanatory variable and the objective variable) and a coefficient of determination (an index representing how well the predicted value of the objective variable matches the actual value of the objective variable) for each of the plurality of generated prediction models.

[0045] (5) Display of learning results Below the learning execution icon 84, a result display section 85 showing the result of machine learning is displayed. The user can confirm the result of machine learning by viewing the display content of the result display section 85. The result display section 85 displays a model name display section 85a, a coefficient display section 85b, a graph display section 85d, and a feature amount display section 85e.

[0046] In the model name display section 85a, the name of the prediction model (algorithm) used for the learning result displayed in the result display section 85 is displayed. By default, in the model name display section 85a, the name of the prediction model with the highest coefficient of determination among the plurality of prediction models (hereinafter also referred to as the "best model") (in the example shown in FIG. 2, "lar") is displayed.

[0047] When the user performs an operation of pressing the pull-down button displayed at the right end of the model name display section 85a, a pick list 85f is displayed in which a plurality of algorithms used for generating the prediction model are arranged in order of accuracy (in descending order of the coefficient of determination). By selecting any one of the plurality of prediction models (algorithms) displayed in the pick list 85f, the user can change the learning result displayed in the result display section 85 to the learning result by the prediction model (selected model) generated by the algorithm selected by the user.

[0048] In the coefficient display section 85b, the correlation coefficient and the coefficient of determination of the prediction model displayed in the model name display section 85a are displayed. The graph display section 85d displays a graph representing the analysis result by the prediction model displayed in the model name display section 85a. In the feature amount display section 85e, the ranking of the feature amounts used for the prediction model displayed in the model name display section 85a is displayed.

[0049] <Utilization phase (phase of utilizing the machine learning result)> FIG. 3 is a diagram schematically showing an example of the learning utilization screen 70b displayed on the display device 2. When the user performs an operation of selecting the tab 90, the learning utilization screen 70b shown in FIG. 3 is displayed on the screen of the display device 2. The user can calculate (predict) the value of the objective variable for any value of the explanatory variable using the prediction model generated in the learning phase by simply performing a simple operation according to the display content of the learning utilization screen 70b.

[0050] On the learning utilization screen 70b, an explanatory variable input section 91, an objective variable display section 92, a consideration data display section 93, and a model selection section 94 are displayed.

[0051] In the explanatory variable input section 91, a title column 91a, an input column 91b, a range column 91c, a calculation execution icon 91d, and an initial value icon 91e are displayed.

[0052] The title column 91a, the input column 91b, and the range column 91c are arranged and displayed in this order from the left. In the title column 91a, the title names of a plurality of explanatory variables are vertically arranged and displayed, and in the range column 91c, the minimum value and the maximum value of each explanatory variable are displayed. In the input column 91b, the median value of each explanatory variable is input as the initial value. The user can change the value of the explanatory variable in the input column 91b from the initial value to an arbitrary value within the range displayed in the range column 91c. Note that FIG. 3 shows an example in which four items, "Classification B", "Classification a", "Classification e", and "Classification c" among the items displayed in the title column 91a are target variables, and the other items are fixed values.

[0053] In the model selection unit 94, the name of the prediction model used for calculating the target variable is displayed. The best model is displayed by default in the model selection unit 94. When the user performs an operation of pressing the pull-down button displayed at the right end of the model selection unit 94, a pick list in which the names of a plurality of prediction models generated in the learning phase are arranged in order of accuracy is displayed. The user can change the prediction model used for calculating the target variable to the prediction model selected by the user by selecting one prediction model from the pick list.

[0054] When the user performs an operation of moving the cursor to the calculation execution icon 91d and clicking it (hereinafter also referred to as "calculation execution operation"), the learning device 6 uses the prediction model displayed in the model selection unit 94 to calculate the value of the objective variable for the value of the explanatory variable input in the input field 91b (hereinafter also referred to as "calculation process"). The calculated value of the objective variable is displayed on the objective variable display unit 92. By repeating the operation of changing the value of the explanatory variable input in the input field 91b and recalculating the value of the objective variable, the user can grasp how much the objective variable changes with respect to the change of the explanatory variable. For example, when the explanatory variable is a variable related to the material properties of the battery and the objective variable is a variable related to the storage characteristics of the battery, the user can predict how much the storage characteristics of the battery change with respect to the change of the material properties of the battery. The material properties of the battery are, for example, the physical properties of the positive electrode, negative electrode, separator and electrolyte, and the material properties of the materials used for these.

[0055] [Flowchart] FIG. 4 is a flowchart showing an outline of the processing procedure executed by the learning device 6 in the learning phase. This flowchart is executed when the learning execution screen 70a shown in FIG. 2 is displayed on the screen of the display device 2.

[0056] First, the learning device 6 determines whether or not the above-described file reading operation has been performed (step S10). If the file reading operation has not been performed (NO in step S10), the learning device 6 repeats the process of step S10 and waits for the file reading operation to be performed.

[0057] If the file reading operation has been performed (YES in step S10), the learning device 6 reads the target file, extracts the explanatory variable candidates and the objective variable candidates from the learning data included in the read target file, and displays the extracted explanatory variable candidates and objective variable candidates on the display device 2 in a manner that can be selected by the user (step S11).

[0058] Next, the learning device 6 determines whether or not the above-described variable selection operation has been performed (step S12). If the variable selection operation has not been performed (NO in step S12), the learning device 6 repeats the process of step S12 and waits for the variable selection operation to be performed.

[0059] If the variable selection operation has been performed (YES in step S12), the learning device 6 determines whether or not the above-described learning execution operation has been performed (step S13). If the learning execution operation has not been performed (NO in step S13), the learning device 6 repeats the process of step S13 and waits for the learning execution operation to be performed.

[0060] If the learning execution operation has been performed (YES in step S13), the learning device 6 executes a learning process of performing machine learning with the above-described content (step S14). Thereafter, the learning device 6 causes the display device 2 to display the result of the learning process (step S15).

[0061] FIG. 5 is a flowchart showing an outline of a processing procedure executed by the learning device 6 in the utilization phase. This flowchart is executed when the learning utilization screen 70b shown in FIG. 3 is displayed on the screen of the display device 2.

[0062] First, the learning device 6 determines whether or not the above-described calculation execution operation has been performed (step S20). If the calculation execution operation has not been performed (NO in step S20), the learning device 6 repeats the process of step S20 and waits for the calculation execution operation to be performed.

[0063] If the calculation execution operation has been performed (YES in step S20), the learning device 6 performs the above-described calculation process (step S21). Specifically, the learning device 6 calculates the value of the objective variable with respect to the value of the explanatory variable input in the input field 91b using the prediction model displayed in the model selection unit 94.

[0064] Next, the learning device 6 causes the display device 2 to display the result of the calculation process (step S22).

[0065] As described above, the machine learning system 1 according to this embodiment includes a display device 2 and a learning device 6. The learning device 6 causes the display device 2 to display the result of extracting explanatory variable candidates and objective variable candidates from a target file specified by the user, and analyzes the relationship between the explanatory variables and objective variables selected by the user from among the explanatory variable candidates and objective variable candidates displayed on the display device 2 by machine learning, and causes the display device 2 to display the result.

[0066] According to the machine learning system 1 according to this embodiment, just by the user specifying a target file, candidates for explanatory variables and objective variables included in the target file are displayed on the display device 2. The user can cause the display device 2 to display the analysis result by machine learning by performing a simple operation of selecting desired variables from among the explanatory variable candidates and objective variable candidates displayed on the display device 2. Therefore, the user can perform machine learning with a simple operation without performing complicated operations such as coding.

[0067] Furthermore, the learning device 6 according to this embodiment, in the learning phase, performs machine learning on the explanatory variables and objective variables selected by the user according to a plurality of algorithms to generate a plurality of prediction models for predicting the objective variables from the explanatory variables respectively, calculates the coefficient of determination of each of the generated plurality of prediction models, and causes the display device 2 to display the prediction model with the highest coefficient of determination as the best model. Thereby, the user can know the best model with a simple operation.

[0068] Furthermore, the learning device 6 according to this embodiment, in the learning phase, causes the display device 2 to display a plurality of algorithms, and causes the display device 2 to display the information of the best model or the information of a selected model generated by the algorithm selected by the user from among the plurality of algorithms. Thereby, the user can change the information of the prediction model to be displayed on the display device 2 from the information of the best model to the information of the selected model by performing a simple operation of selecting any one of the plurality of algorithms displayed on the display device 2.

[0069] Furthermore, in the utilization phase, the learning device 6 according to the present embodiment acquires the explanatory variables input by the user, and causes the display device 2 to display the result of calculating the objective variable corresponding to the explanatory variables input by the user using the best model or the selected model. Thereby, the user can predict the value of the objective variable corresponding to any explanatory variable.

[0070] Furthermore, in the learning phase, the machine learning system 1 according to the present embodiment acquires the division ratio of the training data and the test data requested by the user, divides the data included in the target file into the training data and the test data at the acquired division ratio, performs machine learning, and causes the display device 2 to display the result. Thereby, the user can arbitrarily determine the division ratio of the training data and the test data.

[0071] The embodiments disclosed this time should be considered as illustrative in all respects and not restrictive. The scope of the present disclosure is shown not by the above description but by the claims, and it is intended that all modifications within the meaning and scope equivalent to the claims are included.

Description of Reference Numerals

[0072] 1 Machine learning system, 2 Display device, 3 Keyboard, 4 Mouse, 5 Removable disk, 6 Learning device, 10 Support device, 11 Arithmetic unit, 12 Memory unit, 13 Learning data, 15 Display interface, 16 Peripheral device interface, 17 Reading unit, 18 Communication unit, 60 Analysis device, 61 Prediction model, 70a Learning execution screen, 70b Learning utilization screen, 80, 90 Tabs, 81a File selection icon, 81b Path display section, 81c Icon, 81d Status bar, 82 Split selection section, 82a Method selection section, 82b Ratio selection section, 83 Variable selection section, 83a Objective variable selection section, 83b Explanatory variable selection section, 83c Selected variable display section, 84 Learning execution icon, 85 Result display section, 85a Model name display section, 85b Coefficient display section, 85d Graph display section, 85e Feature quantity display section, 85f Pick list, 91 Explanatory variable input section, 91a Title bar, 91b Input field, 91c Range field, 91d Calculation execution icon, 91e Initial value icon, 92 Objective variable display section, 93 Consideration data display section, 94 Model selection section.

Claims

1. A display device and, a learning device connected to the display device, wherein the learning device causes the display device to display the result of extracting candidate explanatory variables that can be selected as explanatory variables and candidate target variables that can be selected as target variables from a target file specified by a user, and causes the display device to display the result of analyzing the relationship between the explanatory variables and the target variables selected by the user from among the candidate explanatory variables and the candidate target variables displayed on the display device by machine learning. A machine learning system.

2. The learning device generates a plurality of prediction models for predicting a target variable from an explanatory variable by performing machine learning on the explanatory variable and the target variable selected by the user according to a plurality of algorithms, respectively, calculates the coefficient of determination of each of the generated plurality of prediction models, and causes the display device to display the prediction model having the highest coefficient of determination as the best model. The machine learning system according to claim 1.

3. The learning device displays the plurality of algorithms on the display device, and causes the display device to display the information of the best model or the information of a selection model that is a prediction model generated by an algorithm selected by the user from among the plurality of algorithms displayed on the display device. The machine learning system according to claim 2.

4. The learning device acquires the explanatory variable input by the user, and causes the display device to display the result of calculating the target variable corresponding to the explanatory variable input by the user using the best model or the selection model. The machine learning system according to claim 3.

5. The learning device acquires the division ratio of training data and test data requested by the user, divides the data included in the target file into training data and test data at the acquired division ratio, and causes the display device to display the result of performing the machine learning. The machine learning system according to any one of claims 1 to 4.

6. The explanatory variable is a variable related to the material of the battery, and the target variable is a variable related to the characteristics of the battery. The machine learning system according to any one of claims 1 to 4.

7. displaying, on a display device, the result of extracting candidate explanatory variables that can be selected as explanatory variables and candidate target variables that can be selected as target variables from a target file specified by a user; A machine learning method including a step of causing the display device to display a result of analyzing, by machine learning, the relationship between an explanatory variable and an objective variable selected by a user from among the explanatory variable candidates and the objective variable candidates displayed on the display device.

Images