Information processing device, information processing method, and program

Abstract

To evaluate the significance of regression analysis results of material sample data about each explanatory variable combination.SOLUTION: An information processing device comprises: a first evaluation unit which performs regression analysis about each explanatory variable combination on the basis of first data containing a plurality of explanatory variables, which are a feature amount, and an objective variable, which is performance, about a plurality kinds of material samples, and evaluates an error for the regression analysis results; a second evaluation unit which performs regression analysis about each explanatory variable combination on the basis of the second data obtained by changing a value of the objective variable in the first data and evaluates an error for the regression analysis results; a generation unit which generates a distribution representing the number of explanatory variable combinations per error for the regression analysis results in the first data and generates a distribution representing the number of explanatory variable combinations serving as an error per error for the regression analysis results in the second data; and an output unit which outputs the comparison results of the generated distributions.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to an information processing device, an information processing method, and a program. [Background technology]

[0002] Patent Document 1 discloses a prediction device that visualizes whether a predicted value is reliable using an index called reliability. The prediction device includes an input unit for inputting explanatory variables and a tolerance, a regression prediction model database for storing regression prediction models, a quantile regression model database for storing quantile regression models, a prediction unit for predicting a dependent variable based on the explanatory variables and the regression prediction model, and a reliability calculation unit for predicting a predicted quantile value based on the explanatory variables and the quantile regression model, calculating an allowable error range from the dependent variable and the allowable error, and calculating the reliability of the predicted dependent variable from the relationship between the allowable error range and the predicted quantile value. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2022-14618 A Summary of the Invention [Problem to be solved by the invention]

[0004] The prediction device of Patent Document 1 does not take into consideration the evaluation of the significance of a regression model. Furthermore, when performing regression analysis for each combination of at least one explanatory variable among a plurality of explanatory variables, it is required to evaluate the significance of each regression analysis result.

[0005] The present disclosure aims to evaluate the significance of the results of the regression analysis of the data of the material samples for each combination of explanatory variables. [Means for solving the problem]

[0006] The information processing device according to the first aspect includes: a first evaluation unit that performs regression analysis for each combination of at least one or more explanatory variables among a plurality of explanatory variables based on first data including a plurality of explanatory variables that are feature quantities of the material samples and a target variable that is performance of the material samples, and evaluates an error for a result of the regression analysis for the combination; a second evaluation unit that performs regression analysis for each combination of at least one or more explanatory variables among the plurality of explanatory variables based on second data in which a value of the target variable in the first data for the plurality of types of material samples is changed, and evaluates an error for a result of the regression analysis for the combination; a generation unit that generates a distribution representing the number of combinations of the explanatory variables that result in the error for each error based on a result of the error evaluation for the result of the regression analysis in the first data, and generates a distribution representing the number of combinations of the explanatory variables that result in the error for each error based on a result of the error evaluation for the result of the regression analysis in the second data; and an output unit that outputs a result of comparing the generated distributions.

[0007] In the information processing device according to the first aspect, a first evaluation unit performs a regression analysis for each combination of at least one explanatory variable among a plurality of explanatory variables based on first data including a plurality of explanatory variables that are feature quantities of the material samples and a response variable that is the performance of the material samples for a plurality of types of material samples, and evaluates an error in the result of the regression analysis for the combination. Here, the "regression analysis" refers to finding a regression coefficient for each explanatory variable when expressing the value of the response variable with the value of the explanatory variable. Also, the "error in the result of the regression analysis" refers to the error between the value of the response variable estimated using the value of the explanatory variables and the regression coefficient of each explanatory variable, and the true value of the response variable.

[0008] Then, a second evaluation unit performs regression analysis for each combination of at least one explanatory variable among the plurality of explanatory variables based on second data obtained by changing the value of the objective variable in the first data for the plurality of types of material samples, and evaluates an error in the result of the regression analysis for the combination.

[0009] Then, a generating unit generates a distribution representing the number of combinations of the explanatory variables that result in the error for each error based on the evaluation result of the error for the result of the regression analysis for the first data, and generates a distribution representing the number of combinations of the explanatory variables that result in the error for each error based on the evaluation result of the error for the result of the regression analysis for the second data. An output unit outputs a result of comparing the generated distributions. This makes it possible to evaluate the significance of the result of the regression analysis of the material sample data for each combination of explanatory variables.

[0010] The information processing device of the second aspect, in the information processing device of the first aspect, further includes a determination unit that determines the significance of a result of the regression analysis on the first data based on a result of comparing the generated distributions, and a visualization unit that visualizes, for each of the explanatory variables, the magnitude of the regression coefficient obtained as a result of the regression analysis for each of the combinations of the explanatory variables when it is determined that the result of the regression analysis on the first data is significant.

[0011] In the information processing device according to the second aspect, a determination unit determines the significance of the result of the regression analysis on the first data based on a result of comparing the generated distributions. Then, when a visualization unit determines that the result of the regression analysis on the first data is significant, the visualization unit visualizes the magnitude of the regression coefficient obtained as a result of the regression analysis for each of the combinations of the explanatory variables for each of the explanatory variables. This allows a user to grasp the explanatory variables with large regression coefficients in the result of the regression analysis for each combination of the explanatory variables.

[0012] An information processing device according to a third aspect is the information processing device according to the second aspect, further including a reception unit that receives a selection of at least one explanatory variable, and the first evaluation unit further performs the regression analysis on a combination of the selected one or more explanatory variables and evaluates an error in a result of the regression analysis for the combination.

[0013] In the information processing device according to the third aspect, the receiving unit receives the selection of at least one explanatory variable. Then, the first evaluation unit performs the regression analysis on a combination of the selected one or more explanatory variables and evaluates an error in a result of the regression analysis on the combination. This makes it possible to perform a regression analysis on a desired combination of explanatory variables, facilitating feature selection.

[0014] An information processing method according to a fourth aspect includes the steps of: performing a regression analysis for each combination of at least one or more explanatory variables among a plurality of explanatory variables based on first data for a plurality of types of material samples, the first data including a plurality of explanatory variables that are feature quantities of the material samples and a target variable that is performance of the material samples; performing a regression analysis for each combination of at least one or more explanatory variables among the plurality of explanatory variables based on second data in which a value of the target variable in the first data for the plurality of types of material samples is changed; evaluating an error in a result of the regression analysis for the combination; generating a distribution representing the number of combinations of the explanatory variables that result in the error for each error based on a result of the error evaluation for the result of the regression analysis for the first data; generating a distribution representing the number of combinations of the explanatory variables that result in the error for each error based on a result of the error evaluation for the result of the regression analysis for the second data; and outputting a result of comparing the generated distributions.

[0015] In the information processing method according to the fourth aspect, a regression analysis is performed for each combination of at least one explanatory variable among the multiple explanatory variables based on first data including multiple explanatory variables that are feature quantities of the material samples and a response variable that is the performance of the material samples, and an error is evaluated for the result of the regression analysis for the combination. A regression analysis is performed for each combination of at least one explanatory variable among the multiple explanatory variables based on second data in which the value of the response variable in the first data for the multiple types of material samples is changed, and an error is evaluated for the result of the regression analysis for the combination. Then, based on the error evaluation result for the regression analysis result for the first data, a distribution representing the number of combinations of the explanatory variables that result in the error is generated for each error, and based on the error evaluation result for the regression analysis result for the second data, a distribution representing the number of combinations of the explanatory variables that result in the error is generated for each error. Then, a result of comparing the generated distributions is output. This makes it possible to evaluate the significance of the result of the regression analysis of the data of the material samples for each combination of the explanatory variables.

[0016] A program according to a fifth aspect is a program for causing a computer to execute a process of performing a regression analysis for each combination of at least one or more explanatory variables among a plurality of explanatory variables based on first data including a plurality of explanatory variables that are feature quantities of the material samples and a target variable that is performance of the material samples, and evaluating an error for a result of the regression analysis for the combination; performing a regression analysis for each combination of at least one or more explanatory variables among the plurality of explanatory variables based on second data in which a value of the target variable in the first data for the plurality of types of material samples is changed, and evaluating an error for a result of the regression analysis for the combination; generating a distribution representing the number of combinations of the explanatory variables that result in the error for each error based on a result of the error evaluation for the result of the regression analysis for the first data; generating a distribution representing the number of combinations of the explanatory variables that result in the error for each error based on a result of the error evaluation for the result of the regression analysis for the second data; and outputting a result of comparing the generated distributions.

[0017] In the program according to the fifth aspect, a computer performs a regression analysis for each combination of at least one explanatory variable among the multiple explanatory variables based on first data including multiple explanatory variables that are feature quantities of the material samples and a target variable that is the performance of the material samples, and evaluates an error for the result of the regression analysis for the combination. The computer performs a regression analysis for each combination of at least one explanatory variable among the multiple explanatory variables based on second data in which the value of the target variable in the first data for the multiple types of material samples is changed, and evaluates an error for the result of the regression analysis for the combination. Then, the computer generates a distribution representing the number of combinations of the explanatory variables that result in the error for each error based on the evaluation result of the error for the result of the regression analysis for the first data, and generates a distribution representing the number of combinations of the explanatory variables that result in the error for each error based on the evaluation result of the error for the result of the regression analysis for the second data. Then, the computer outputs a result of comparing the generated distributions. This makes it possible to evaluate the significance of the result of the regression analysis of the data of the material samples for each combination of the explanatory variables. Effect of the Invention

[0018] As described above, the present disclosure has the advantage of being able to evaluate the significance of the results of regression analysis of material sample data for each combination of explanatory variables. [Brief description of the drawings]

[0019] [Figure 1] 1 is a schematic block diagram of an information processing system according to an embodiment. [Diagram 2] FIG. 13 is a diagram illustrating an example of a graph showing the number of combinations of explanatory variables that result in each error. [Diagram 3] FIG. 13 is a diagram illustrating an example of a graph showing the number of combinations of explanatory variables that result in each error. [Figure 4]FIG. 13 is a diagram showing an example of visualizing the magnitude of the regression coefficient for each explanatory variable. [Diagram 5] FIG. 2 is a diagram illustrating an example of the configuration of a cloud server and a computer of a user terminal according to the embodiment. [Figure 6] 5 is a flowchart showing an example of an information processing routine performed by the information processing device according to the embodiment. [Figure 7] 5 is a flowchart showing an example of an information processing routine performed by the information processing device according to the embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0020] Hereinafter, an information processing system according to an embodiment will be described with reference to the drawings.

[0021] Fig. 1 is a block diagram showing an example of a functional configuration of an information processing system 10 according to an embodiment. As shown in Fig. 1, the information processing system 10 includes a plurality of user terminals 14A, 14B, ..., 14N and a cloud server 12 which is an example of an information processing device. The plurality of user terminals 14A, 14B, ..., 14N and the cloud server 12 are connected via a network 16 such as the Internet. In the following, when referring to one user terminal, it is simply referred to as a user terminal 14.

[0022] (User terminal) Each of the user terminals 14A to 14N transmits to the cloud server 12 measurement data relating to the material sample measured by a plurality of analysis methods.

[0023] Each of the multiple user terminals 14A, 14B, . . . , 14N is operated by a multiple different users.

[0024] The user inputs measurement data on the material sample to be analyzed into the user terminal 14 that the user operates. The measurement data on the material sample to be analyzed is, for example, data measured by X-ray diffraction method, small angle X-ray scattering method, data measured by a microscope, data measured by Raman spectroscopy, and data measured by infrared spectroscopy.

[0025] (Cloud Server) Cloud server 12 stores the measurement data of a plurality of material samples, and analyzes each of the plurality of material samples from the measurement data using an analysis method, and stores analysis data representing the analysis results. For example, the material sample is analyzed from the measurement data using an analysis method for the measurement data, such as an analysis method for X-ray diffraction, an analysis method for small-angle X-ray scattering, an analysis method for a microscope image, an analysis method for Raman spectroscopy, an analysis method for infrared spectroscopy, or the like.

[0026] The cloud server 12 performs regression analysis for each combination of at least one of the multiple explanatory variables based on data obtained from the measurement data and analysis data for multiple types of material samples, the multiple explanatory variables being characteristic quantities of the material samples, and the target variable being the performance of the material samples, and evaluates the results of the regression analysis.

[0027] Specifically, as shown in FIG. 1, the cloud server 12 has the functions of an acquisition unit 20, a first evaluation unit 22, a second evaluation unit 24, a generation unit 26, an output unit 28, a judgment unit 30, a visualization unit 32, and a reception unit 34, and a database 36.

[0028] The acquiring unit 20 acquires measurement data relating to the material sample measured by a plurality of analysis methods from each of the plurality of user terminals 14A-14N, and stores the data in the database 36. The acquiring unit 20 analyzes the material sample from the measurement data using the analysis method, and stores analysis data representing the analysis results in the database 36.

[0029] Moreover, the acquiring unit 20 acquires, from the database 36, first data including a plurality of explanatory variables which are feature quantities of the material samples and a response variable which is the performance of the material samples, for a plurality of types of material samples.

[0030] Specifically, a plurality of feature amounts and performance are obtained from each measurement data and each analysis data for the material sample, and the plurality of feature amounts are used as a plurality of explanatory variables and the performance is used as a response variable to obtain first data.

[0031] The acquiring unit 20 acquires second data in which the value of the objective variable is changed in the first data for the multiple types of material samples. Specifically, the value of the objective variable of the first data is exchanged between the material samples to acquire second data for the multiple types of material samples, the second data including multiple explanatory variables that are feature quantities of the material samples and objective variables that are performances of the material samples.

[0032] The first evaluation unit 22 performs regression analysis for each combination of at least one explanatory variable among the multiple explanatory variables based on the first data for multiple types of material samples, and evaluates the error in the result of the regression analysis for the combination.

[0033] Specifically, the first evaluation unit 22 performs a regression analysis for each combination of at least one explanatory variable among the multiple explanatory variables based on the first data for the multiple types of material samples, thereby obtaining a regression coefficient for each explanatory variable when expressing the value of the objective variable with the value of the explanatory variable.Then, the first evaluation unit 22 obtains an error between the value of the objective variable estimated by the value of the explanatory variable and the regression coefficient of each explanatory variable obtained as a result of the regression analysis, and the true value of the objective variable, for each combination of at least one explanatory variable among the multiple explanatory variables, based on the first data for the multiple types of material samples, thereby evaluating the error in the result of the regression analysis for the combination.

[0034] The second evaluation unit 24, similar to the first evaluation unit 22, performs regression analysis for each combination of at least one explanatory variable among the multiple explanatory variables based on the second data for multiple types of material samples, and evaluates the error in the results of the regression analysis for the combination.

[0035] The generating unit 26 generates, for each error, a distribution representing the number of combinations of explanatory variables that result in the error, based on an evaluation result of the error for the result of the regression analysis on the first data for each combination of explanatory variables. The generating unit 26 generates, for each error, a distribution representing the number of combinations of explanatory variables that result in the error, based on an evaluation result of the error for the result of the regression analysis on the second data for each combination of explanatory variables.

[0036] The output unit 28 generates a screen showing the results of comparing the generated distributions, as shown in FIGS. 2 and 3, and displays the screen on the user terminal 14.

[0037] Figures 2 and 3 are examples of graphs showing the number of combinations of explanatory variables that result in each cross validation error (CVE) for a given CVE, in order to allow comparison between the results of regression analysis using the first data and the results of regression analysis using the second data.

[0038] Figure 2 shows an example where the result of regression analysis on the first data shows that there are more combinations of explanatory variables with small cross-validation errors than combinations of explanatory variables with large cross-validation errors. Figure 2 also shows an example where the result of regression analysis on the second data shows that there is no difference between the number of combinations of explanatory variables with large cross-validation errors and the number of combinations of explanatory variables with small cross-validation errors.

[0039] FIG. 3 shows an example in which the result of the regression analysis on the first data has more combinations of explanatory variables with small cross-validation errors than the number of combinations of explanatory variables with large cross-validation errors. FIG. 3 also shows an example in which the result of the regression analysis on the second data has more combinations of explanatory variables with small cross-validation errors than the number of combinations of explanatory variables with large cross-validation errors. In other words, FIG. 3 shows an example in which the graph showing the number of combinations of explanatory variables that result in each cross-validation error does not differ between the result of the regression analysis on the first data and the result of the regression analysis on the second data.

[0040] The determination unit 30 determines the significance of the result of the regression analysis on the first data based on the result of comparing the generated distributions. For example, the absolute value of the difference in the number of combinations of explanatory variables, which is each cross-validation error, between the result of the regression analysis on the first data and the result of the regression analysis on the second data is calculated, and if the sum of the absolute values ​​of the differences is equal to or greater than a threshold, it is determined that there is a difference in the generated distribution between the result of the regression analysis on the first data and the result of the regression analysis on the second data, and that the result of the regression analysis on the first data is significant.

[0041] When it is determined that the result of the regression analysis on the first data is significant, the visualization unit 32 generates a screen that visualizes the magnitude of the regression coefficient of each explanatory variable obtained as a result of the regression analysis on the first data for each combination of explanatory variables, as shown in FIG. 4, and displays the screen on the user terminal 14.

[0042] FIG. 4 shows an example in which the magnitude of the regression coefficient of each explanatory variable obtained as a result of regression analysis on the first data for each combination of explanatory variables is color-coded and visualized for each explanatory variable. Also, in FIG. 4, the left side shows the results of regression analysis for combinations of explanatory variables with smaller cross-validation errors, and the right side shows the results of regression analysis for combinations of explanatory variables with larger cross-validation errors. Also, FIG. 4 shows an example in which the regression coefficients are color-coded to distinguish whether they are positively correlated or negatively correlated.

[0043] The receiving unit 34 receives the selection of at least one explanatory variable from the user terminal 14 on the screen that visualizes the magnitude of the regression coefficient. In the above FIG. 4, an example is shown in which three explanatory variables are selected (see the three dotted frames).

[0044] The first evaluation unit 22 again performs regression analysis on the combination of one or more selected explanatory variables, evaluates the error in the result of the regression analysis for the combination, and displays a screen showing the error evaluation results on the user terminal 14.

[0045] The user terminal 14 and the cloud server 12 can be realized by, for example, a computer 50 as shown in Fig. 5. The computer 50 realizing the user terminal 14 and the cloud server 12 includes a CPU 51, a memory 52 as a temporary storage area, and a non-volatile storage unit 53. The computer 50 also includes an input / output interface (I / F) 54 to which an input / output device or the like (not shown) is connected, and a read / write (R / W) unit 55 that controls reading and writing of data from and to a recording medium 59. The computer also includes a network I / F 56 that is connected to a network such as the Internet. The CPU 51, the memory 52, the storage unit 53, the input / output I / F 54, the R / W unit 55, and the network I / F 56 are connected to one another via a bus 57.

[0046] The storage unit 53 can be realized by a hard disk drive (HDD), a solid state drive (SSD), a flash memory, etc. The storage unit 53 as a storage medium stores a program for causing the computer to function. The CPU 51 reads the program from the storage unit 53, expands it in the memory 52, and sequentially executes the processes contained in the program.

[0047] Next, the operation of the information processing system 10 according to the embodiment will be described.

[0048] When measurement data regarding the material sample is input to the user terminal 14, the measurement data regarding the material sample is transmitted to the cloud server 12. When the measurement data regarding the material sample is transmitted from the user terminal 14 to the cloud server 12, the cloud server 12 stores the measurement data regarding the material sample in the database 36. In this way, the measurement data regarding a plurality of material samples is stored in the database 36.

[0049] In addition, the cloud server 12 analyzes each of the multiple material samples from the measurement data stored in the database 36 using an analysis method, obtains analysis data representing the analysis results, and stores the analysis data in the database 36.

[0050] When a request for analyzing the data of a material sample is input to the user terminal 14, the request for analyzing the data of the material sample is transmitted to the cloud server 12. Then, the cloud server 12 executes the information processing routine shown in FIGS.

[0051] In step S100, the acquiring unit 20 acquires, from the database 36, first data for a plurality of types of material samples, including a plurality of explanatory variables which are feature quantities of the material samples and a response variable which is the performance of the material samples.

[0052] In step S102, the first evaluation unit 22 sets one combination of at least one explanatory variable among the multiple explanatory variables as a processing target.

[0053] In step S104, the first evaluation unit 22 performs regression analysis for each combination of explanatory variables to be processed based on first data including multiple explanatory variables that are feature quantities of the material samples and a target variable that is the performance of the material samples for multiple types of material samples.

[0054] In step S106, the first evaluation unit 22 evaluates an error in the result of the regression analysis for the combination of explanatory variables to be processed.

[0055] In step S108, the first evaluation unit 22 judges whether or not the processes of steps S102 to S106 have been executed for all combinations of explanatory variables. If there is a combination of explanatory variables for which the processes of steps S102 to S106 have not been executed, the process returns to step S102 and sets the combination of explanatory variables. On the other hand, if the processes of steps S102 to S106 have been executed for all combinations of explanatory variables, the process proceeds to step S110.

[0056] In step S110, the acquiring unit 20 acquires second data for the multiple types of material samples, the second data being obtained by changing the value of the objective variable in the first data for the multiple types of material samples.

[0057] In step S112, the second evaluation unit 24 sets one combination of at least one explanatory variable among the multiple explanatory variables as a processing target.

[0058] In step S114, the second evaluation unit 24 performs a regression analysis on the combination of explanatory variables to be processed, based on the second data on the multiple types of material samples.

[0059] In step S116, the second evaluation unit 24 evaluates an error in the result of the regression analysis for the combination of explanatory variables to be processed.

[0060] In step S118, the second evaluation unit 24 judges whether or not the processes of steps S112 to S116 have been executed for all combinations of explanatory variables. If there is a combination of explanatory variables for which the processes of steps S112 to S116 have not been executed, the process returns to step S112 and sets the combination of explanatory variables. On the other hand, if the processes of steps S112 to S116 have been executed for all combinations of explanatory variables, the process proceeds to step S120.

[0061] In step S120, the generating unit 26 generates, for each error, a distribution representing the number of combinations of explanatory variables that result in the error, based on the evaluation result of the error in the first data.

[0062] In step S122, the generating unit 26 generates, for each error, a distribution representing the number of combinations of explanatory variables that result in the error, based on the evaluation result of the error in the second data.

[0063] In step S124, the output unit 28 outputs the result of comparing the generated distributions.

[0064] In step S126, the determination unit 30 determines whether the result of the regression analysis on the first data is significant based on the result of comparing the generated distributions. If the result of the regression analysis on the first data is not significant, the information processing routine is terminated. On the other hand, if the result of the regression analysis on the first data is significant, the process proceeds to step S128.

[0065] In step S128, the visualization unit 32 generates a screen that visualizes the magnitude of the regression coefficient of each explanatory variable obtained as a result of the regression analysis on the first data for each combination of explanatory variables, and displays the screen on the user terminal 14.

[0066] The user looks at the screen and selects explanatory variables to be used when modeling the performance, which is the target variable.

[0067] In step S130, the receiving unit 34 determines whether or not the selection of at least one explanatory variable has been received from the user terminal 14 on the screen that visualizes the magnitude of the regression coefficient. If the selection of an explanatory variable has been received from the user terminal 14, the process proceeds to step S132.

[0068] In step S132, the first evaluation section 22 performs a regression analysis on the combination of one or more selected explanatory variables.

[0069] In step S134, the first evaluation unit 22 evaluates the error for the result of the regression analysis for that combination, and causes the evaluation result to be displayed on the user terminal 14, thereby terminating the information processing routine.

[0070] As described above, the cloud server of the information processing system according to the embodiment generates a distribution representing the number of combinations of explanatory variables that result in errors for each error based on an error evaluation result for the results of regression analysis of a first data set including a plurality of explanatory variables that are feature quantities of the material samples and a response variable that is the performance of the material samples for a plurality of types of material samples, generates a distribution representing the number of combinations of explanatory variables that result in errors for each error based on an error evaluation result for the results of regression analysis of a second data set in which the value of the response variable in the first data for a plurality of types of material samples is changed, and outputs a result of comparing the generated distributions. This makes it possible to evaluate the significance of the results of the regression analysis of the data of the material samples for each combination of explanatory variables.

[0071] Furthermore, the significance of the result of the regression analysis on the first data is determined based on the result of comparing the generated distributions, and when it is determined that the result of the regression analysis on the first data is significant, the magnitude of the regression coefficient obtained as a result of the regression analysis for each combination of explanatory variables is visualized for each explanatory variable, thereby allowing the user to grasp the explanatory variables with large regression coefficients in the result of the regression analysis for each combination of explanatory variables.

[0072] In addition, the selection of at least one explanatory variable is accepted, a regression analysis is performed for a combination of the selected one or more explanatory variables, and an error is evaluated for the result of the regression analysis for the combination. This makes it possible to perform a regression analysis for a desired combination of explanatory variables, facilitating feature selection.

[0073] In the above embodiment, the processes performed by each device have been described as software processes performed by executing a program, but they may be hardware processes. Alternatively, they may be a combination of both software and hardware processes. Furthermore, the programs stored in the ROM may be stored in various storage media and distributed.

[0074] In addition, although the second data is obtained by exchanging the values ​​of the objective variables of the first data between material samples, the present invention is not limited to this example. For example, the second data may be obtained by any other method as long as the value of the objective variable in the first data is changed.

[0075] Furthermore, the present invention is not limited to the above, and it goes without saying that various modifications can be made without departing from the spirit and scope of the present invention. [Explanation of symbols]

[0076] 10. Information Processing Systems 12 Cloud server (information processing device) 14 User terminal 20 Acquisition Department 22 First Evaluation Section 24 Second Evaluation Section 26 Generation part 28 Output section 30 Judgment section 32 Visualization part 34 Reception Department 50 Computer 52 Memory 53 Storage section

Claims

1. a first evaluation unit that performs a regression analysis for each combination of at least one explanatory variable among a plurality of explanatory variables based on first data including a plurality of explanatory variables that are feature amounts of the material samples and a response variable that is performance of the material samples for a plurality of types of material samples, and evaluates an error in a result of the regression analysis for the combination; a second evaluation unit that performs a regression analysis for each combination of at least one explanatory variable among the plurality of explanatory variables based on second data obtained by changing a value of the objective variable in the first data for the plurality of types of material samples, and evaluates an error in a result of the regression analysis for the combination; generating a distribution representing the number of combinations of the explanatory variables resulting in the error for each error based on an evaluation result of the error for the result of the regression analysis on the first data; a generation unit that generates, for each error, a distribution representing the number of combinations of the explanatory variables that result in the error, based on an evaluation result of the error for the result of the regression analysis on the second data; an output unit that outputs a result of comparing the generated distributions; An information processing device comprising:

2. a determination unit that determines the significance of the result of the regression analysis on the first data based on a result of comparing the generated distributions; a visualization unit that visualizes, for each of the explanatory variables, a magnitude of a regression coefficient obtained as a result of the regression analysis in each of the combinations of the explanatory variables when it is determined that the result of the regression analysis in the first data is significant; The information processing apparatus according to claim 1 , further comprising:

3. A reception unit that receives a selection of at least one explanatory variable, The information processing apparatus according to claim 2 , wherein the first evaluation unit further performs the regression analysis on a combination of the one or more selected explanatory variables, and evaluates an error in a result of the regression analysis on the combination.

4. A computer-implemented method for detecting a plurality of performing a regression analysis for each combination of at least one explanatory variable among the plurality of explanatory variables based on first data including a plurality of explanatory variables that are feature quantities of the material samples and a response variable that is performance of the material samples for a plurality of types of material samples, and evaluating an error in a result of the regression analysis for the combination; performing a regression analysis for each combination of at least one explanatory variable among the plurality of explanatory variables based on second data obtained by changing a value of the objective variable in the first data for the plurality of types of material samples, and evaluating an error in a result of the regression analysis for the combination; generating a distribution representing the number of combinations of the explanatory variables that result in the error for each error based on an evaluation result of the error for the result of the regression analysis on the first data; generating a distribution representing the number of combinations of the explanatory variables that result in the error for each error based on an evaluation result of the error for the result of the regression analysis on the second data; outputting a result of comparing the generated distributions; An information processing method comprising:

5. performing a regression analysis for each combination of at least one explanatory variable among the plurality of explanatory variables based on first data including a plurality of explanatory variables that are feature quantities of the material samples and a response variable that is performance of the material samples for a plurality of types of material samples, and evaluating an error in a result of the regression analysis for the combination; performing a regression analysis for each combination of at least one explanatory variable among the plurality of explanatory variables based on second data obtained by changing the value of the objective variable in the first data for the plurality of types of material samples, and evaluating an error in a result of the regression analysis for the combination; generating a distribution representing the number of combinations of the explanatory variables resulting in the error for each error based on an evaluation result of the error for the result of the regression analysis on the first data; A distribution is generated based on an evaluation result of the error for the result of the regression analysis on the second data, the distribution representing the number of combinations of the explanatory variables that result in the error for each error. Output the results of comparing the generated distributions A program that causes a computer to execute a process.

Images