Information processing device, information processing method, and program

Abstract

This information processing device (1) comprises: an extraction unit (11) that extracts information of at least a portion of values from a result of regression analysis of a data set that is used for learning; and an output unit (12) that outputs information of a step in which data corresponding to the information of the portion of values extracted by the extraction unit (11) was acquired from among steps in which the data set was acquired.

Description

Information processing device, information processing method, and program

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

[0002] Regression analysis is a method for predicting deterioration of an object of analysis, such as a system or device. Regression analysis analyzes the relationship between one or more explanatory variables and one target variable in a data set, and predicts unknown data using the relationship resulting from the analysis. For example, an anomaly detection method described in Patent Document 1 acquires data from multiple sensors, divides the trajectory of the data space into multiple clusters based on changes in the data over time, models the divided clusters using regression analysis, and calculates outliers as anomaly candidates.

[0003] JP 2010-92355 A

[0004] The conventional technology described in Patent Document 1 has a problem in that the interpretability of regression-analyzed data is low. For example, when regression-analyzed data is used for learning degradation detection, variability in the learning data can cause overdetection. When data interpretability is low, it is difficult to distinguish between normal data and variability data because the source of the variability data is unknown. For this reason, there is a possibility that learning data containing a large amount of variability will be used for learning.

[0005] The present disclosure is intended to solve the above-mentioned problems, and aims to provide an information processing device that can provide information that improves the interpretability of a regression-analyzed dataset.

[0006] The information processing device according to the present disclosure includes an extraction unit that extracts information on at least some values ​​from the results of regression analysis of a dataset used for learning, and an output unit that outputs information on a process from which data corresponding to the information on the some values ​​extracted by the extraction unit was obtained, among the processes from which the dataset was obtained.

[0007] According to the information processing device according to the present disclosure, it is possible to provide information that improves the interpretability of a regression-analyzed dataset.

[0008] 1 is a block diagram showing an example of a configuration of an information processing device according to embodiment 1. FIG. 3A is a diagram showing results of regression analysis of a dataset. FIG. 3C is a diagram showing results of regression analysis of various datasets. FIG. 4 is a flowchart showing an information processing method according to embodiment 1. FIG. 4B is a diagram showing an overview of an outlier extraction process. FIG. 5A is a diagram showing results of regression analysis of a dataset in which outliers have been processed. FIG. 5C is a diagram showing an overview of a section extraction process. FIG. 6A is a block diagram showing an example of a hardware configuration for realizing the functions of the information processing device according to embodiment 1. FIG. 6B is a block diagram showing an example of a configuration of an information processing device according to embodiment 6. FIG. 7C is a flowchart showing an information processing method according to embodiment 7. FIG. 7B is a block diagram showing an example of a configuration of an information processing device according to embodiment 7. FIG. 8 is a flowchart showing an information processing method according to embodiment 8. FIG. 8B is a block diagram showing an example of a configuration of an information processing device according to embodiment 8. FIG. 9 is a flowchart showing an information processing method according to embodiment 9. FIG. 9 is a block diagram showing a configuration of a first variant of an information processing device according to embodiment 9. FIG. 10 is a flowchart showing a second variant of an information processing method according to embodiment 10. FIG. 11 is a diagram showing an overview of a process of presenting data of degradation candidates. FIG. 12 is a block diagram showing an example of a configuration of an information processing device according to embodiment 10. FIG. 13 is a flowchart showing an information processing method according to embodiment 10.

[0009] Embodiment 1 (Basic Configuration of Information Processing Device) Fig. 1 is a block diagram showing an example configuration of an information processing device 1 according to embodiment 1. In Fig. 1, the information processing device 1 extracts information on at least some values ​​from the results of regression analysis of a dataset used for learning (hereinafter referred to as regression analysis results), and outputs information on processes from which data corresponding to the extracted information on some values ​​were obtained, among the processes from which the dataset was obtained. The some values ​​extracted from the regression analysis results are values ​​of analysis points obtained by regression analysis of the data. The data corresponding to the information on some values ​​is data before regression analysis that corresponds to the analysis points, which are some values.

[0010] For example, the information processing device 1 may acquire regression analysis results from an external device via a network and generate process information based on the acquired information. The external device may be a regression analysis device that performs regression analysis on a preset data set. Alternatively, the information processing device 1 may acquire regression analysis results stored in its own storage unit and generate process information based on the acquired information.

[0011] (Process Information) Process information is information indicating a process for which data corresponding to information on a portion of values ​​extracted from the regression analysis results has been obtained. A "process" is, for example, a process represented by parameters that characterize the repetition of a work procedure or the operation of an object of regression analysis, for which data corresponding to information on the portion of values ​​has been obtained.

[0012] (Obtaining Regression Analysis Results) The regression analysis results include, for example, information indicating the attributes and characteristics of the dataset, information indicating an overview of the regression model, information indicating coefficients and their meanings, information indicating the goodness of fit of the regression model, information indicating statistical significance, information indicating the reliability of the prediction, and information indicating the constraints and assumptions of the regression model. The information indicating the attributes and characteristics of the dataset includes, for example, parameters that serve as explanatory variables and response variables, and information indicating the source of the dataset. The information indicating the overview of the regression model includes, for example, whether the regression model used is a simple regression analysis or a multiple regression analysis, and an explanation of the formula and parameters that represent the regression model. The information indicating the coefficients and their meanings is information indicating the coefficients of each explanatory variable and their meanings, and indicates, for example, which explanatory variables affect the response variable. The information indicating the goodness of fit of the regression model includes, for example, evaluation information on how well the regression model fits the data, such as the coefficient of determination or the distribution of residuals. The information indicating statistical significance includes, for example, evaluation information on whether each coefficient is statistically significant. The information indicating the reliability of the prediction includes, for example, information indicating a prediction interval or a confidence interval. The information indicating the constraints and assumptions of the regression model includes, for example, assumptions such as linearity, independence, and homoscedasticity, as well as information regarding the scope of application of the regression analysis.

[0013] For example, let us consider the acquisition of regression analysis results detected by detecting deterioration of a railway motor. A vibration sensor attached to a railway vehicle detects vibrations generated in the railway motor while the railway vehicle is in operation. Vibration data sequentially detected by the vibration sensor constitutes a data set to be subjected to regression analysis and is output to a regression analysis device (not shown in FIG. 1 ). Furthermore, operation data synchronized with the vibration data is output to the regression analysis device and the information processing device 1. The operation data is data indicating the operation of the railway vehicle powered by the railway motor. Furthermore, synchronization with the vibration data refers to temporal synchronization. For example, operation data indicating an operation performed at a certain time is synchronized with vibration data generated at that time. The regression analysis device performs regression analysis using the operation data as an explanatory variable and the vibration data as a response variable. The regression analysis results include, for example, parameters serving as explanatory variables and response variables, and information indicating a confidence interval. The regression analysis results generated by the regression analysis device are output from the regression analysis device to the information processing device 1. The dataset to be subjected to regression analysis may be, for example, data obtained by sequentially detecting the rotational speed of a railway motor, or data obtained by sequentially detecting the position, speed, acceleration, or torque of a robot axis of an industrial robot, or data obtained by sequentially detecting vibrations generated in a motor provided in an elevator, or data obtained by sequentially detecting the power used in one opening and closing of a platform door installed on a station platform.

[0014] (Variation in Regression Analysis Results) FIG. 2 is a diagram showing the results of regression analysis, specifically, the results of regression analysis detected in railway motor deterioration detection. In FIG. 2, the horizontal axis represents railway vehicle operation data, which is the explanatory variable, and the vertical axis represents vibration data, which is the dependent variable. C1 represents the confidence interval. Point D is a point within the confidence interval C1 among the points obtained by regression analysis of vibration data when the railway motor is in a normal state (hereinafter referred to as normal data). Predicted value E is the predicted value of the dependent variable obtained by regression analysis. Point F is a point outside the confidence interval C1 among the points obtained by regression analysis of normal data. Here, the presence of a point corresponding to normal data outside the confidence interval C1 is referred to as variability in normal data. Generally, when deterioration detection training is performed using normal data with variability as training data, the detection accuracy of the trained model obtained by this training decreases. In particular, detection using a regression model loses time series information, making it difficult to fully analyze outliers and more susceptible to overdetection.

[0015] Therefore, the information processing device 1 extracts information on at least some of the values ​​from the regression analysis results, and outputs information on the processes for which data corresponding to the extracted information on some of the values ​​was obtained, among the processes for which a data set was obtained. By referring to the process information, it is possible to identify the processes for which data corresponding to the information on some of the values ​​extracted from the regression analysis results was obtained. Based on the identified processes, even if there is variation in the data corresponding to some of the values, it is possible to distinguish the data from other data.

[0016] For example, the information processing device 1 presents, as process information, operation data that is an explanatory variable that yields point F outside the confidence interval C1 among points obtained by regression analysis of normal data. By referencing this process information, data with variability can be removed from the learning data. Furthermore, for example, the information processing device 1 can automatically extract tasks or driving sections in which data variability occurs continuously as process information using a parameter (e.g., kilometers) that characterizes the repetition of driving patterns and a distance standard specified by the user, and display this on the display device 2. By referencing the display information, data with variability can be removed from the learning data.

[0017] The display device 2 displays information output from the information processing device 1. For example, the display device 2 acquires and displays information on processes in which data corresponding to information on some values ​​extracted by the extraction unit was acquired from the processes in which a dataset was acquired from the information processing device 1. By referring to the process information displayed on the display device 2, it is possible to identify the process in which data corresponding to information on some values ​​extracted from the data resulting from the regression analysis was acquired. Based on the identified process, even if there is variation in the data corresponding to some values, it is possible to distinguish the data from other data. Note that the display device 2 may be a display unit included in the information processing device 1 or a display unit included in an external device connected to and communicating with the information processing device 1 via a network. For example, the display device 2 is an LCD (Liquid Crystal Display) or an organic EL (Electroluminescence) display.

[0018] The input device 3 accepts input operations from outside the information processing device 1. When the information processing device 1 is a smartphone or a tablet terminal, the input device 3 is, for example, a touch panel that is integrated with the screen of the display device 2. When the information processing device 1 is a PC, the input device 3 is, for example, a mouse or a keyboard.

[0019] The information processing device 1 includes an extraction unit 11 and an output unit 12. For example, the information processing device 1 is realized by a computer. A memory included in the computer stores a program constituting an information processing application for realizing each function of the extraction unit 11 and the output unit 12. A processor included in the computer reads the information processing application from the memory and executes the information processing application, thereby realizing each function of the extraction unit 11 and the output unit 12.

[0020] (Extraction Unit) The extraction unit 11 extracts information on at least some values ​​from the results of regression analysis of a dataset used for learning. For example, if the information processing device 1 includes a communication unit (not shown in FIG. 1 ) and is communicatively connected to a regression analysis device via a network, the extraction unit 11 controls the communication unit to acquire information indicating the results of regression analysis of the dataset from the regression analysis device via the network. Also, for example, if the information processing device 1 includes a storage unit (not shown in FIG. 1 ) and information indicating the results of regression analysis of the dataset is stored in the storage unit, the extraction unit 11 may read and acquire the information indicating the results of regression analysis of the dataset from the storage unit. Note that in this case, the information processing device 1 does not need to include a communication unit.

[0021] (Output Unit) The output unit 12 outputs information about a process for which data corresponding to information about a partial value extracted by the extraction unit 11 was acquired, among the processes for which a dataset was acquired. For example, a description will be given of a dataset in which vibration data detected from vibrations generated in a railway motor of a railway vehicle while the railway vehicle was repeatedly operated on a certain line is subjected to regression analysis. The extraction unit 11 extracts outliers of the confidence interval from the results of regression analysis of the dataset. The outliers are the partial values ​​extracted from the regression analysis results by the extraction unit 11. The output unit 12 uses a kilometer distance, which is a parameter characterizing the repetition of the driving pattern, and a distance standard specified by the user, to output information about a driving section expressed in kilometers, in which outlier points extracted by the extraction unit 11 occur consecutively and data corresponding to the outliers were acquired. This information about the driving section is information about the process. Here, the distance standard is a standard in which a distance of less than 0.5 km is considered close and a distance of more than 0.5 km is considered far.

[0022] For example, the output unit 12 calculates an invariant parameter Inv(t) for each repetitive operation of the railway vehicle based on control information for controlling the operation of the railway vehicle. Here, the parameter Inv(t) satisfies the condition (1) that it monotonically increases with time t and the condition (2) that at least I exists that is greater than or equal to 0 and that the range from 0 to I is repeated for each lap. The above-mentioned kilometer distance is the parameter Inv(t). Note that the parameter Inv(t) is not limited to kilometer distance, and various parameters depending on the subject of the regression analysis may be used. For example, if the subject of the regression analysis is a series of tasks performed by an industrial robot shown in FIG. 3B (described later), the task progress indicating how far the series of tasks has been completed may be used as the parameter Inv(t). Furthermore, if the subject of the regression analysis is the opening and closing of platform doors shown in FIG. 3C (described later), the power consumed when the platform doors open and close once may be used as the parameter Inv(t). The output unit 12 may acquire control information of the object of regression analysis (e.g., a railway vehicle), calculate a cumulative function of all parameters Inv(t) based on the acquired control information, and search for parameters that satisfy the above conditions (1) and (2) among the parameters for which the cumulative functions have been calculated, thereby automatically calculating the parameter Inv(t) indicating the kilometer distance. Furthermore, the parameter Inv(t) may be set in the output unit 12 by a user using the input device 3.

[0023] The dataset may be a dataset obtained by repeatedly performing a process. Figures 3A, 3B, and 3C are diagrams showing outliers extracted from the results of regression analysis of various datasets obtained by repeatedly performing a process. Figure 3A is a diagram showing the relationship between outliers corresponding to vibration data of a railway motor detected when a railway vehicle repeatedly travels around a circular route and the kilometers representing the route on which the railway vehicle travels, over time. In Figure 3A, if a portion of the route causes an outlier to occur, the outliers for each lap appear consecutively at kilometers corresponding to that portion. Figure 3B is a diagram showing the relationship between outliers in vibration data of a robot axis detected when an industrial robot repeatedly performs a series of tasks and task progress representing the progress of the series of tasks, over time. Similarly, in Figure 3B, if a portion of the series of tasks causes an outlier to occur, the outliers for each lap appear consecutively at task progress corresponding to that portion. Figure 3C is a diagram showing the relationship between outliers in power usage data when platform doors open and close in response to the arrival and departure of railway vehicles at the platform, and the kilometers representing the route on which the railway vehicle travels, over time. In Figure 3C, if an outlier occurs during either the opening or closing of the platform doors, the outliers for each opening and closing of the platform doors will appear consecutively during either the opening or closing operation. By performing regression analysis on the data set obtained by repeatedly performing the process, it is possible to understand how the data changes over time. Furthermore, if the data set is repeatedly obtained over a long period of time, it is possible to understand long-term trends and patterns and use them to make future predictions.

[0024] Next, an information processing method according to the first embodiment will be described. FIG. 4 is a flowchart showing the information processing method according to the first embodiment. The extraction unit 11 extracts information on at least some values ​​from the results of regression analysis of a dataset used for learning (step ST1). The output unit 12 outputs information on processes for which datasets have been acquired, for which data corresponding to some of the values ​​extracted by the extraction unit 11 have been acquired (step ST2). For example, the output unit 12 outputs display information for displaying the process information to the display device 2. After step ST2, the display device 2 displays the process information based on the display information from the output unit 12. By executing the series of processes shown in FIG. 4, the information processing device 1 can provide process information, which is information that improves the interpretability of a regression-analyzed dataset.

[0025] (Variation 1) The extraction unit 11 extracts information on some values ​​determined based on the predicted value of the dependent variable for the explanatory variables. For example, the extraction unit 11 calculates the importance of each explanatory variable using a tree-based model such as a decision tree or a random forest, and determines which explanatory variables are important for prediction based on the calculated importance. Feature quantities with high importance are likely to provide important information for predicting the dependent variable. Furthermore, for example, the extraction unit 11 may quantify the contribution of each explanatory variable to the prediction of the dependent variable as a SHAP (Shapley Additive Explanations) value, and extract information on some values ​​that contribute little to the prediction by the regression model based on the SHAP value. This enables the information processing device 1 to extract information on some values ​​determined based on the predicted value of the dependent variable from the regression analysis results.

[0026] Furthermore, the extraction unit 11 may extract information on some values ​​from the results of regression analysis of the dataset using a machine learning model that, when a dataset is input, outputs information on some values ​​determined based on predicted values ​​of the objective variable for the explanatory variables. The machine learning model may be stored in a storage unit (not shown in FIG. 1 ) included in the information processing device 1, or may be stored in an external storage device that is provided separately from the information processing device 1 and can be connected to the information processing device 1 for communication. In this case, the extraction unit 11 controls the communication unit to establish a communication connection with the external storage device. The dataset transmitted from the extraction unit 1 via the communication unit is input to the machine learning model stored in the external storage device, and information output from the machine learning model is received by the communication unit and acquired by the extraction unit 11 or the output unit 12.

[0027] (Variation 2) Furthermore, the extraction unit 11 may extract information on some values ​​that fall outside the confidence interval that includes the predicted value E of the dependent variable for the explanatory variables. For example, the extraction unit 11 checks whether the value of each point in the regression analysis result shown in FIG. 2 is within the confidence interval C1. If the value of point F falls outside the confidence interval C1, the extraction unit 11 detects point F as an outlier. An outlier may be determined based on how far the point F is from the upper or lower limit of the confidence interval. The extraction unit 11 extracts information on the value of point F that falls outside the confidence interval C1. This enables the information processing device 1 to extract information on some values ​​that fall outside the confidence interval C1 that includes the predicted value E of the dependent variable from the regression analysis result.

[0028] (Variation 3) The extraction unit 11 may extract information on a portion of values ​​determined by a user from the results of regression analysis of a dataset. For example, the user sets the value of a specific explanatory variable to the extraction unit 11 using the input device 3. The extraction unit 11 extracts information on a portion of values ​​corresponding to the value of the explanatory variable set by the user from the regression analysis results. This enables the information processing device 1 to extract information on a portion of values ​​determined by the user from the regression analysis results.

[0029] (Variation 4) Furthermore, the extraction unit 11 may extract information on some values ​​indicating values ​​outside the confidence interval determined by the user, which includes the predicted value of the objective variable for the explanatory variables. For example, the output unit 12 displays the regression analysis result shown in FIG. 2 on the display device 2. The user, having reference to the information displayed on the display device 2, sets a new confidence interval using the input device 3. The extraction unit 11 extracts information on some values ​​indicating values ​​outside the confidence interval determined by the user. This enables the information processing device 1 to extract information on some values ​​indicating values ​​outside the confidence interval determined by the user from the regression analysis result.

[0030] (Variation 5) The output unit 12 may also output, to the display device 2, information on a process in which data corresponding to information on the partial values ​​extracted by the extraction unit 11 was acquired. FIG. 5 is a diagram illustrating an outline of an outlier extraction process. For example, when the extraction unit 11 acquires the regression analysis result shown in FIG. 2 , it extracts information on the value of point F that is outside the confidence interval C1 in the regression model. The output unit 12 generates display information for displaying a diagram (1) in which the information on the value of point F extracted by the extraction unit 11 is plotted in a space with axes representing kilometers over time and point values ​​resulting from the regression analysis of the vibration data, as shown in FIG. 5 , and outputs the generated display information to the display device 2. Based on this display information, the display device 2 displays a diagram (1) in which outlier points obtained each time the railway vehicle travels around the track are plotted, as shown in FIG. 3A . The outliers are the partial values ​​extracted from the regression analysis result by the extraction unit 11. Furthermore, data corresponding to the outliers in the dataset to be regression analyzed are data corresponding to information on the partial values. By referring to Figure (1), it is possible to recognize outliers that appear in multiple rotations.

[0031] The output unit 12 identifies the kilometres in which outliers appear consecutively for each lap in the relationship shown in Figure (1), and extracts the section represented by the identified kilometres. Information about this section is information about the process in which data corresponding to the information about the partial value extracted by the extraction unit 11 was acquired. The output unit 12 generates display information for displaying Figure (2) showing the section in which outlier points appear consecutively, and Figure (2) is displayed on the display device 2 based on this display information. By referring to Figure (2), it is possible to recognize the section in which outlier points appear consecutively.

[0032] (Variation 6) The output unit 12 may also output to the display device 2 information indicating the relationship between process information in which data corresponding to the partial value information extracted by the extraction unit 11 was acquired and the partial value information. As described above, in FIG. 5 , the outliers are the partial values ​​extracted from the regression analysis results by the extraction unit 11, and the data corresponding to the outliers in the dataset to be regression analyzed is the data corresponding to the partial value information. The output unit 12 identifies a portion B corresponding to a section in which outlier points appear consecutively on a route map A on which the railway vehicle travels, generates display information for displaying a diagram (3) in which the identified portion B is distinguished from other route portions, and outputs the generated display information to the display device 2. Based on this display information, the display device 2 displays a diagram (3) in which the section in which the outlier points appear consecutively and the portion B on the route map A are linked. Here, the information indicating the section in which the outlier points appear consecutively is the process information, and the information indicating the portion B on the route map A is the partial value information. By referring to the diagram (3), it is possible to recognize which portion of the route map the outlier points appear consecutively. For example, by referring to FIG. (3), the cause of the deviation, such as a winding of the track in portion B, may be inferred, and data corresponding to the outliers acquired in portion B may be processed. Furthermore, by instructing the information processing device 1 using the input device 3, data corresponding to the outliers acquired in portion B may be removed from the target of regression analysis. Note that the output unit 12 may display any one of FIG. (1), FIG. (2), and FIG. (3) shown in Modifications 5 and 6 on the display device 2, or may display any combination of these.

[0033] FIG. 6 is a diagram showing the results of regression analysis after outliers have been removed. In FIG. 6, the horizontal axis represents railway vehicle operation data, which is the explanatory variable, and the vertical axis represents vibration data, which is the dependent variable. Confidence interval C2 is a confidence interval set by the regression analysis results after removing outlier point F shown in FIG. 2. Point D is one of the points obtained by regression analysis of normal data that is included in confidence interval C2. Predicted value E is a predicted value of the dependent variable obtained by regression analysis. Point F is one of the points obtained by regression analysis of normal data that falls outside confidence interval C2. Comparing FIG. 2 with FIG. 6, it can be seen that removing the data corresponding to the outlier points obtained in part B significantly reduces the number of points that fall outside confidence interval C2.

[0034] (Interval Extraction) Next, details of the process of extracting intervals in which outlier points appear consecutively will be described. For example, the extraction unit 11 sets consecutive intervals based on the size of the intervals between values ​​in a process among values ​​obtained by regression analysis of a dataset, and extracts information on values ​​included in the set intervals. This makes it possible to extract intervals in which values ​​appear consecutively. FIG. 7A is a diagram showing outlier points plotted on the kilometer axis. FIG. 7B is a diagram showing intervals extracted using a coarse distance criterion that includes outlier points plotted on the kilometer axis. FIG. 7C is a diagram showing intervals extracted using a fine distance criterion that includes outlier points plotted on the kilometer axis.

[0035] In the example shown in FIG. 7A , in a portion where multiple outlier points appear on the kilometer-distance axis, the extraction unit 11 determines the distance between the most distant outlier point as distance criterion m1 and the distance between the next most distant outlier point on the kilometer-distance axis as distance criterion m2. Note that distance criterion m1 or distance criterion m2 may be set by the user using the input device 3 or may be preset in the extraction unit 11. When extracting a section using distance criterion m1, as shown in FIG. 7B , in a portion where multiple outlier points appear on the kilometer-distance axis, the extraction unit 11 determines that two adjacent points separated by a distance m1 are reference points, and if the outlier data adjacent to these points fall within the distance criterion m1, extracts a range of kilometers including these points as a continuous section G1 in which the outlier points are continuous. On the other hand, if the outlier points fall outside the distance criterion m1, the extraction unit 11 determines that the outlier points are discontinuous. Similarly, when extracting a section using distance criterion m2, as shown in Figure 7C, the extraction unit 11 uses two adjacent points separated by a distance m2 as references in a section where multiple outlier points appear on the kilometer axis. If the adjacent outlier points fall within distance criterion m2, the extraction unit 11 extracts the kilometer range including these outlier points as a continuous section in which the outlier points are continuous. On the other hand, if the adjacent outlier points fall outside distance criterion m2, the extraction unit 11 determines that the outlier points are discontinuous. As a result, in the example of Figure 7C, continuous sections G2 and G3 are extracted.

[0036] Next, a hardware configuration for realizing the functions of the information processing device 1 will be described. Fig. 8 is a block diagram showing the hardware configuration for realizing the functions of the information processing device 1. The functions of the extraction unit 11 and the output unit 12 included in the information processing device 1 are realized by a processing circuit. That is, the information processing device 1 includes a processing circuit for executing the processes of steps ST1 to ST2 shown in Fig. 4. The processing circuit may be a CPU (Central Processing Unit) that executes a program stored in a memory.

[0037] The extraction unit 11 acquires, for example, information indicating the results of regression analysis of a dataset used for learning, which information is received from an external device by a communication unit included in the information processing device 1, via the input interface 100. Furthermore, if information indicating the results of regression analysis of a dataset used for learning is stored in a storage unit included in the information processing device 1, the extraction unit 11 reads and acquires the information from the storage unit via the input interface 100. In this case, the information processing device 1 does not need to be equipped with a communication unit.

[0038] The output unit 12 outputs, for example, via the output interface 101, information on a process in which data corresponding to information on some values ​​extracted by the extraction unit 11 was acquired, among the processes in which a dataset was acquired, to the display device 2. The output unit 12 may also transmit the results of the process to an external device by controlling the communication unit via the output interface 101.

[0039] The functions of the extraction unit 11 and the output unit 12 included in the information processing device 1 are realized by software, firmware, or a combination of software and firmware. The software or firmware is written as a program and stored in the memory 103.

[0040] The processor 102 reads and executes programs stored in the memory 103 to realize the functions of the extraction unit 11 and the output unit 12 included in the information processing device 1. For example, the information processing device 1 includes the memory 103 for storing programs that, when executed by the processor 102, result in the processing of steps ST1 to ST2 shown in FIG. 4 being executed. These programs cause a computer to execute the procedures or methods of the processing performed by the extraction unit 11 and the output unit 12. The memory 103 may be a computer-readable storage medium that stores programs for causing a computer to function as the extraction unit 11 and the output unit 12.

[0041] The memory 103 may be, for example, a non-volatile or volatile semiconductor memory such as a RAM (Random Access Memory), a ROM (Read Only Memory), a flash memory, an EPROM (Erasable Programmable Read Only Memory), or an EEPROM (Electrically-EPROM) (registered trademark), a magnetic disk, a flexible disk, an optical disk, a compact disk, a mini disk, or a DVD.

[0042] Some of the functions of the extraction unit 11 and the output unit 12 included in the information processing device 1 may be realized by dedicated hardware, and the other parts may be realized by software or firmware. For example, the function of the extraction unit 11 may be realized by a processing circuit that is dedicated hardware, and the function of the output unit 12 may be realized by the processor 102 reading and executing a program stored in the memory 103. In this way, the processing circuit can realize the above functions by hardware, software, firmware, or a combination of these.

[0043] As described above, the information processing device 1 according to the first embodiment includes an extraction unit 11 that extracts information on at least some values ​​from the results of regression analysis of a dataset used for learning, and an output unit 12 that outputs information on processes from which the dataset was acquired, at which data corresponding to the information on the some values ​​extracted by the extraction unit 11 was acquired. By referring to the process information, it is possible to identify the process from which data corresponding to the information on the some values ​​extracted from the data resulting from the regression analysis was acquired. Based on the identified process, for example, even if there is variation in the data corresponding to the some values, it is possible to distinguish the data from other data. In this way, the information processing device 1 can provide information that improves the interpretability of a dataset subjected to regression analysis.

[0044] In the information processing device 1 according to the first embodiment, the extraction unit 11 extracts information on some values ​​determined based on the predicted values ​​of the dependent variables for the explanatory variables. This enables the information processing device 1 to extract information on some values ​​determined based on the predicted values ​​of the dependent variables from the regression analysis results.

[0045] In the information processing device 1 according to the first embodiment, the extraction unit 11 extracts information on some values ​​that fall outside the confidence interval that includes the predicted value E of the dependent variable for the explanatory variables. This enables the information processing device 1 to extract, from the regression analysis results, information on some values ​​that fall outside the confidence interval C1 that includes the predicted value E of the dependent variable.

[0046] In the information processing device 1 according to the first embodiment, the extraction unit 11 extracts information on some values ​​determined by the user from the results of regression analysis of a data set. This enables the information processing device 1 to extract information on some values ​​determined by the user from the regression analysis results.

[0047] In the information processing device 1 according to the first embodiment, the extraction unit 11 extracts information on a portion of values ​​that indicate values ​​outside the confidence interval determined by the user and that includes the predicted value of the objective variable for the explanatory variables. This enables the information processing device 1 to extract information on a portion of values ​​that indicate values ​​outside the confidence interval determined by the user from the regression analysis results.

[0048] In the information processing device 1 according to the first embodiment, a data set is acquired by repeatedly performing processes. By performing regression analysis on the data set acquired by repeatedly performing processes, it is possible to understand how the data changes over time. Furthermore, when data sets are acquired repeatedly over a long period of time, it is possible to understand long-term trends and patterns and use them to make future predictions.

[0049] In the information processing device 1 according to the first embodiment, the output unit 12 outputs information on the process in which data corresponding to the information on some values ​​extracted by the extraction unit 11 was obtained to the display device 2. By referring to this display information, it is possible to recognize the process in which data corresponding to the information on some values ​​was obtained.

[0050] In the information processing device 1 according to the first embodiment, the output unit 12 outputs information indicating the relationship between the information on the partial value and the information on the process in which data corresponding to the information on the partial value extracted by the extraction unit 11 was obtained, to the display device 2. By referring to this display information, it is possible to recognize the information indicating the relationship between the information on the partial value and the information on the process in which data corresponding to the information on the partial value was obtained.

[0051] In the information processing device 1 according to the first embodiment, the extraction unit 11 sets consecutive intervals based on the size of intervals between values ​​in a process among values ​​obtained by regression analysis of a data set, and extracts information on values ​​included in the set intervals. This enables the information processing device 1 to extract intervals in which each value appears consecutively.

[0052] The information processing method according to the first embodiment includes a step (ST1) in which the extraction unit 11 extracts information on at least some values ​​from the results of regression analysis of a dataset used for learning, and a step (ST2) in which the output unit 12 outputs information on a process in which, among the processes in which the dataset was acquired, data corresponding to the values ​​extracted by the extraction unit 11 was acquired. In this way, the information processing method according to the first embodiment can provide information that improves the interpretability of the regression-analyzed dataset.

[0053] A computer that executes the program according to the first embodiment executes the following steps: an extraction unit 11 extracts information on at least some values ​​from the results of regression analysis of a dataset used for learning (ST1); and an output unit 12 outputs information on a process, among the processes in which the dataset was acquired, in which data corresponding to the values ​​extracted by the extraction unit 11 was acquired (ST2). In this way, the computer that executes the program according to the first embodiment functions as an information processing device 1, and is capable of providing information that improves the interpretability of a dataset that has been regression analyzed.

[0054] Second Embodiment In the first embodiment, information on the process in which data corresponding to information on a portion of values ​​extracted by the extraction unit was acquired is output. In the second embodiment, however, it is possible to select whether or not to use information on the extracted portion of values ​​for learning.

[0055] FIG. 9 is a block diagram showing an example configuration of an information processing device 1A according to a second embodiment. In FIG. 9, the information processing device 1A extracts information on at least some values ​​from the regression analysis results and outputs information on processes for which data corresponding to the extracted information on some values ​​has been acquired, among the processes for which a data set has been acquired. For example, the information processing device 1A acquires the regression analysis results from an external device via a network and generates process information based on the acquired information. The external device may be a regression analysis device that performs regression analysis on a preset data set. Furthermore, the information processing device 1A may acquire the regression analysis results stored in its own storage unit and generate process information based on the acquired information.

[0056] The learning device 4 performs learning using the data set output from the output unit 12A as learning data. For example, the learning device 4 is a regression analysis device that performs regression analysis using a regression model. Note that while FIG. 9 illustrates a case where the learning device 4 is an external device of the information processing device 1A, the information processing device 1A may also include a learning unit having the functions of the learning device 4.

[0057] Information processing device 1A includes an extraction unit 11A, an output unit 12A, and a selection unit 13. For example, information processing device 1A is realized by a computer. A memory included in the computer stores a program constituting an information processing application for realizing each function of extraction unit 11A, output unit 12A, and selection unit 13. A processor included in the computer executes the information processing application read from the memory, thereby realizing each function of extraction unit 11A, output unit 12A, and selection unit 13.

[0058] (Selection Unit) The selection unit 13 accepts a selection of whether to use information on some values ​​extracted by the extraction unit 11A for learning. For example, assume that a user, referring to diagram (3) in FIG. 5, infers that the cause of the deviation in portion B is a winding track, and determines that using data corresponding to the outliers acquired in portion B for learning will result in overfitting the regression model. In this case, the user uses the input device 3 to input setting information indicating that the data corresponding to the outliers acquired in portion B extracted by the extraction unit 11A will not be used for learning. The selection unit 13 identifies data acquired in portion B based on the setting information input using the input device 3 and associates unusable information with the identified data to instruct it not to be used for learning. The output unit 12A receives the regression analysis results including information on some values ​​associated with unusable information from the extraction unit 11A, generates a dataset excluding data corresponding to the outliers based on the unusable information, and outputs the generated dataset to the learning device 4 as learning data. This enables the learning device 4 to perform learning using learning data with little variance.

[0059] Furthermore, if the user determines that using data corresponding to outliers acquired in part B for learning will not result in overlearning, the user uses the input device 3 to input setting information indicating that the data corresponding to the outliers acquired in part B extracted by the extraction unit 11A may be used for learning. The selection unit 13 identifies the data acquired in part B based on the setting information input using the input device 3 and associates usability information with the identified data to indicate that it may be used for learning. The output unit 12A inputs, from the extraction unit 11A, the regression analysis results including information on some values ​​associated with the usability information, generates a dataset including data corresponding to the outliers based on the usability information, and outputs the generated dataset to the learning device 4 as learning data. This makes it possible to prevent unnecessary reduction in learning data.

[0060] As described above, the information processing device 1A according to the second embodiment includes the selection unit 13 that receives a selection of whether or not to use, for learning, information on some values ​​extracted by the extraction unit 11. By including the selection unit 13, the information processing device 1A can select whether or not to use, for learning, information on some values ​​extracted from the regression analysis results.

[0061] Embodiment 3 In the second embodiment, the selection of whether or not to use the information on some values ​​extracted from the regression analysis results for learning is accepted, but in the third embodiment, if the information on the process from which data corresponding to the information on some values ​​is acquired complies with the set conditions, it is automatically determined that the information on some values ​​will not be used for learning.

[0062] FIG. 10 is a block diagram showing an example configuration of an information processing device 1B according to embodiment 3. In FIG. 10, the information processing device 1B extracts information on at least some values ​​from the regression analysis results and outputs information on processes for which data corresponding to the extracted information on some values ​​has been acquired, among the processes for which a data set has been acquired. For example, the information processing device 1B acquires the regression analysis results from an external device via a network and generates process information based on the acquired information. The external device may be a regression analysis device that performs regression analysis on a preset data set. Alternatively, the information processing device 1B may acquire the regression analysis results stored in its own storage unit and generate process information based on the acquired information.

[0063] Information processing device 1B includes an extraction unit 11, an output unit 12A, and a determination unit 14. For example, information processing device 1B is realized by a computer. A memory included in the computer stores a program constituting an information processing application for realizing the functions of extraction unit 11, output unit 12A, and determination unit 14. A processor included in the computer executes the information processing application read from the memory, thereby realizing the functions of extraction unit 11, output unit 12A, and determination unit 14.

[0064] (Determination Unit) The determination unit 14 determines whether or not information about a process from which data corresponding to information about a portion of the values ​​extracted by the extraction unit 11 has been obtained conforms to a set condition. For example, the determination unit 14 is set with a set condition that a predetermined number of curves exist on a track included within a section range of a certain length, and with route information indicating the shape of the track, and the output unit 12A outputs, to the determination unit 14, information indicating, for example, portion B in diagram (3) shown in FIG. 5 as information about a process indicating a section in which deviating data continuously appears. In this case, the determination unit 14 determines whether or not the track in portion B conforms to the set condition based on a result of comparing the route information with the process information.

[0065] If the track in part B complies with the set conditions, it is estimated that the track in part B is winding, causing outlier points to appear consecutively. If the set conditions are met, the determination unit 14 does not use the values ​​extracted by the extraction unit 11 for learning. For example, the determination unit 14 outputs a data set that does not include data acquired in part B to the learning device 4. This allows the learning device 4 to perform learning using learning data with little variance.

[0066] If the track in part B does not comply with the set conditions, it is assumed that there is no cause of deviation in the track in part B. If the track does not comply with the set conditions, the determination unit 14 may use the value extracted by the extraction unit 11 for learning. For example, the determination unit 14 outputs a data set including the data acquired in part B to the learning device 4. This makes it possible to prevent unnecessary reduction in learning data.

[0067] (Setting Condition) The setting condition is a condition indicating that the process in which data corresponding to information on some values ​​extracted from the regression analysis results was obtained is in a state in which data variation is likely to occur. For example, the setting condition may indicate a change in the external environment, non-uniformity in structure or material, mechanical wear or deterioration, differences in operators or operating techniques, or changes in measurement conditions or environmental conditions, which are expected to be factors that are likely to cause data variation.

[0068] Next, an information processing method according to the third embodiment will be described. FIG. 11 is a flowchart showing the information processing method according to the third embodiment. The extraction unit 11 extracts information on at least some values ​​from the results of regression analysis of a dataset used for learning (step ST1A). The output unit 12A outputs information on processes from which data corresponding to the values ​​extracted by the extraction unit 11 were obtained, among the processes from which datasets were obtained, to the determination unit 14 (step ST2A). The determination unit 14 determines whether the process information output from the extraction unit 11 complies with the set conditions (step ST3A). If the set conditions are complied with, the determination unit 14 does not use the values ​​extracted by the extraction unit 11 for learning. On the other hand, if the set conditions are not complied with, the determination unit 14 may use the values ​​extracted by the extraction unit 11 for learning. By executing the series of processes shown in FIG. 11 , the information processing device 1B can provide process information, which improves the interpretability of a regression-analyzed dataset, and can accurately determine whether to use information on some values ​​extracted from the regression analysis results for learning.

[0069] As described above, the information processing device 1B according to the third embodiment includes the determination unit 14 that determines whether or not information on a process from which data corresponding to information on some values ​​extracted by the extraction unit 11 has been obtained complies with the set conditions. If the information complies with the set conditions, the value extracted by the extraction unit 11 is not used for learning. This allows the information processing device 1B to automatically determine whether or not to use data corresponding to information on some values ​​extracted from the regression analysis results for learning.

[0070] Fourth Embodiment In the third embodiment, when process information from which data corresponding to information on some values ​​is acquired complies with a set condition, it is automatically determined that the information on some values ​​is not to be used for learning. However, in the fourth embodiment, the information on some values ​​that is not to be used for learning or the data corresponding to the information on some values ​​that is not to be used for learning is labeled.

[0071] FIG. 12 is a block diagram showing an example configuration of an information processing device 1C according to embodiment 4. In FIG. 12, the information processing device 1C extracts information on at least some values ​​from the regression analysis results and outputs information on processes for which data corresponding to the extracted information on some values ​​has been acquired, among the processes for which a data set has been acquired. For example, the information processing device 1C acquires the regression analysis results from an external device via a network and generates process information based on the acquired information. The external device may be a regression analysis device that performs regression analysis on a preset data set. Furthermore, the information processing device 1C may acquire the regression analysis results stored in its own storage unit and generate process information based on the acquired information.

[0072] The information processing device 1C includes an extraction unit 11, an output unit 12A, a determination unit 14, and a labeling unit 15. For example, the information processing device 1C is realized by a computer. A memory included in the computer stores a program constituting an information processing application for realizing each function of the extraction unit 11, the output unit 12A, the determination unit 14, and the labeling unit 15. A processor included in the computer executes the information processing application read from the memory, thereby realizing each function of the extraction unit 11, the output unit 12A, the determination unit 14, and the labeling unit 15.

[0073] (Labeling Unit) The labeling unit 15 labels information on some values ​​extracted by the extraction unit 11 and not used for learning, or data corresponding to information on some values ​​extracted by the extraction unit 11 and not used for learning. For example, the labeling unit 15 assigns label information indicating that the point F, which is extracted by the extraction unit 11 and falls outside the confidence interval C1 in the regression model shown in Fig. 2, will not be used as learning data. Alternatively, the labeling unit 15 may assign label information indicating that the data before regression analysis will not be used as learning data, which corresponds to the point F, which falls outside the confidence interval C1 in the regression model shown in Fig. 2.

[0074] Next, an information processing method according to the fourth embodiment will be described. FIG. 13 is a flowchart showing the information processing method according to the fourth embodiment. The extraction unit 11 extracts at least some value information from the results of regression analysis of a dataset used for learning (step ST1B). The determination unit 14 determines whether the process information output from the extraction unit 11 complies with the set conditions (step ST2B). If the set conditions are complied with, the determination unit 14 does not use the values ​​extracted by the extraction unit 11 for learning. Alternatively, if the set conditions are not complied with, the determination unit 14 may use the values ​​extracted by the extraction unit 11 for learning. Based on the determination result by the determination unit 14, the labeling unit 15 labels the part of value information not used for learning or data corresponding to the part of value information not used for learning (step ST3B). Based on the label information, the output unit 12A identifies the part of value information not used for learning or data corresponding to the part of value information not used for learning, and outputs a dataset not including data corresponding to the part of value information not used for learning to the learning device 4 (step ST4B). Furthermore, similar to the first embodiment, the output unit 12A may output to the display device 2 display information for displaying information about processes for which data corresponding to some values ​​extracted by the extraction unit 11 have been acquired, among the processes for which a dataset has been acquired. The display device 2 displays the information about the processes for which data corresponding to some values ​​have been acquired, based on the display information. By executing the series of processes shown in FIG. 13 , the information processing device 1C can provide process information, which improves the interpretability of a regression-analyzed dataset. Furthermore, the information processing device 1C can accurately distinguish, based on label information, data corresponding to some value information extracted from the regression analysis results and not used for learning.

[0075] (Variation) Values ​​extracted by the extraction unit 11 may be learned using a different learning method than values ​​not extracted. For example, values ​​not extracted from the regression analysis results by the extraction unit 11 are considered to be values ​​obtained by regression analysis, and values ​​extracted from the regression analysis results by the extraction unit 11 are considered to be values ​​obtained by k-nearest neighbor algorithm. Generally, regression analysis is effective for data sets that do not deviate significantly from normal values ​​and have a cohesive distribution. This is because regression analysis is based on the assumption that there is a linear relationship between the explanatory variables and the objective variable. When data do not deviate significantly from normal values ​​and are cohesive, a linear relationship is often observed, and the data can be appropriately modeled in regression analysis. Data corresponding to values ​​not extracted from the regression analysis results by the extraction unit 11 can be considered to have a cohesive distribution with no significant outliers. Therefore, for example, when the learning unit 16 identifies a value not extracted from the regression analysis results by the extraction unit 11 based on the label information assigned by the labeling unit 15, it performs regression analysis on the data set corresponding to the identified value.

[0076] On the other hand, when a new data point is given, the k-nearest neighbor method classifies the data point into the class that appears most frequently among the k nearest training data points. Because the k-nearest neighbor method does not have a complex model, it is possible to model nonlinear relationships. Data corresponding to values ​​extracted from the regression analysis results by the extraction unit 11 can be said to have a large number of outliers and a disorganized distribution. Therefore, for example, when the learning unit 16 determines that a value is extracted from the regression analysis results by the extraction unit 11 based on the label information assigned by the labeling unit 15, it performs k-nearest neighbor learning on the data set corresponding to the determined value. This allows the information processing device 1C to adopt different learning algorithms depending on the trends in the data.

[0077] Although the case where regression analysis and k-nearest neighbor method are used has been shown, the present invention is not limited to these as long as the learning method is in accordance with the trend of the data.

[0078] As described above, the information processing device 1C according to the fourth embodiment includes the labeling unit 15 that labels information on some values ​​that are extracted by the extraction unit 11 and not used for learning, or data corresponding to information on some values ​​that are extracted by the extraction unit 11 and not used for learning. Based on the label information, the information processing device 1C can accurately distinguish data corresponding to information on some values ​​that are extracted from the regression analysis results and not used for learning.

[0079] In the information processing device 1C according to the fourth embodiment, values ​​extracted by the extraction unit 11 are learned using a different learning method from that used for values ​​not extracted. This enables the information processing device 1C to adopt different learning algorithms depending on the trends in the data.

[0080] Fifth Embodiment In the second to fourth embodiments, an information processing device connected to a learning device has been described, but in the fifth embodiment, an information processing device including a learning unit will be described.

[0081] FIG. 14 is a block diagram showing an example configuration of an information processing device 1D according to embodiment 5. In FIG. 14 , the information processing device 1D extracts information on at least some values ​​from the regression analysis results and outputs information on processes for which data corresponding to the extracted information on some values ​​has been acquired, among the processes for which a data set has been acquired. For example, the information processing device 1D acquires the regression analysis results from an external device via a network and generates process information based on the acquired information. The external device may be a regression analysis device that performs regression analysis on a preset data set. Furthermore, the information processing device 1D may acquire regression analysis results stored in its own storage unit and generate process information based on the acquired information.

[0082] The information processing device 1D includes an extraction unit 11, an output unit 12, a determination unit 14, and a learning unit 16. For example, the information processing device 1D is realized by a computer. A memory included in the computer stores a program constituting an information processing application for realizing each function of the extraction unit 11, the output unit 12, the determination unit 14, and the learning unit 16. A processor included in the computer executes the information processing application read from the memory, thereby realizing each function of the extraction unit 11, the output unit 12, the determination unit 14, and the learning unit 16.

[0083] (Learning Unit) The learning unit 16 performs learning based on a data set excluding values ​​extracted by the extraction unit 11, based on settings made by the user or the determination result made by the determination unit 14. For example, the learning unit 16 identifies information about some values ​​not used for learning or data corresponding to information about some values ​​not used for learning based on the determination result made by the determination unit 14, and performs learning such as degradation detection using the data set excluding the identified data as learning data. Also, for example, when the information processing device 1D is connected to the input device 3 shown in FIG. 1 , the user sets values ​​not to be used for learning among the values ​​extracted by the extraction unit 11 using the input device 3. For example, the values ​​not to be used for learning are set by linking information using the input device 3. The learning unit 16 performs learning using the data set excluding data set set by the user as learning data.

[0084] Next, an information processing method according to the fifth embodiment will be described. FIG. 15 is a flowchart showing the information processing method according to the fifth embodiment. The extraction unit 11 extracts information on at least some values ​​from the results of regression analysis of a dataset used for learning (step ST1C). The determination unit 14 determines whether the process information output from the extraction unit 11 complies with the set conditions (step ST2C). If the set conditions are complied with, the determination unit 14 does not use the values ​​extracted by the extraction unit 11 for learning. Alternatively, if the set conditions are not complied with, the determination unit 14 may use the values ​​extracted by the extraction unit 11 for learning. For example, the learning unit 16 performs learning based on a dataset excluding the values ​​extracted by the extraction unit 11, based on the determination result by the determination unit 14 (step ST3C). The output unit 12 outputs the learning results by the learning unit 16 (step ST4C). For example, upon receiving the regression analysis results from the learning unit 16, the output unit 12 generates display information for displaying the regression analysis results and outputs the generated display information to the display device 2. The display device 2 displays the results of the regression analysis performed by the learning unit 16 based on the display information. As in the first embodiment, the output unit 12 may cause the display device 2 to display information about processes for which data corresponding to some of the values ​​extracted by the extraction unit 11 have been obtained, among the processes for which datasets have been obtained. By executing the series of processes shown in Fig. 15, the information processing device 1D can provide process information, which improves the interpretability of a regression-analyzed dataset. Furthermore, the information processing device 1D can perform learning using learning data with little variance.

[0085] (Variation 1) FIG. 16 is a block diagram showing the configuration of an information processing device 1E, which is Variation 1 of the information processing device 1D. In FIG. 16, the information processing device 1E extracts information on at least some values ​​from the regression analysis results and outputs information on processes for which data corresponding to the extracted information on some values ​​has been acquired, among the processes for which a data set has been acquired. For example, the information processing device 1E acquires the regression analysis results from an external device via a network and generates process information based on the acquired information. The external device may be a regression analysis device that performs regression analysis on a preset data set. Alternatively, the information processing device 1E may acquire regression analysis results stored in its own storage unit and generate process information based on the acquired information.

[0086] The information processing device 1E includes an extraction unit 11, an output unit 12, a determination unit 14, a learning unit 16, and a deterioration determination unit 17. For example, the information processing device 1E is realized by a computer. A memory included in the computer stores a program constituting an information processing application for realizing each of the functions of the extraction unit 11, the output unit 12, the determination unit 14, the learning unit 16, and the deterioration determination unit 17. A processor included in the computer executes the information processing application read from the memory, thereby realizing each of the functions of the extraction unit 11, the output unit 12, the determination unit 14, the learning unit 16, and the deterioration determination unit 17.

[0087] (Deterioration Determination Unit) The deterioration determination unit 17 performs deterioration determination based on a trained model trained by the learning unit 16. For example, the trained model trained by the learning unit 16 is a machine learning model that outputs a deterioration determination result by regression analysis when a dataset is input. The deterioration determination result is data indicating whether the result of regression analysis of data is data generated by deterioration of the object of analysis (hereinafter referred to as deteriorated data). For example, deteriorated data is data that deviates significantly from a value corresponding to normal data by exceeding an allowable threshold. The deterioration determination unit 17 performs deterioration determination by inputting a dataset to the trained model trained by the learning unit 16.

[0088] Furthermore, the deterioration determination unit 17 may identify a process in which only outliers corresponding to varied deterioration data appear by comparing a process in which normal data but outliers outside the confidence interval appear (for example, a section of a railroad line) with a process in which outliers corresponding to deterioration data appear. This allows obtaining information on processes in which only deterioration data falls outside the confidence interval, thereby reducing not only overdetection but also overlooking of deterioration.

[0089] Next, a first variation of the information processing method according to the fifth embodiment will be described. FIG. 17 is a flowchart illustrating the first variation of the information processing method according to the fifth embodiment. The extraction unit 11 extracts information on at least some values ​​from the results of regression analysis of a data set used for learning (step ST1D). The determination unit 14 determines whether the process information output from the extraction unit 11 complies with the set conditions (step ST2D). If the process information complies with the set conditions, the determination unit 14 does not use the values ​​extracted by the extraction unit 11 for learning. If the process information does not comply with the set conditions, the determination unit 14 may use the values ​​extracted by the extraction unit 11 for learning. The learning unit 16 performs learning based on a data set excluding the values ​​extracted by the extraction unit 11, for example, based on the determination result by the determination unit 14 (step ST3D). The deterioration determination unit 17 performs deterioration determination based on the trained model trained by the learning unit 16 (step ST4D). The output unit 12 outputs the deterioration determination result by the deterioration determination unit 17 (step ST5D). For example, upon acquiring a degradation assessment result from the degradation assessment unit 17, the output unit 12 generates display information for displaying the degradation assessment result and outputs the generated display information to the display device 2. The display device 2 displays the degradation assessment result from the degradation assessment unit 17 based on the display information. Note that, similar to the first embodiment, the output unit 12 may cause the display device 2 to display information on a process from which data corresponding to some of the values ​​extracted by the extraction unit 11 was acquired, among the processes from which a dataset was acquired. By executing the series of processes shown in FIG. 17 , the information processing device 1E can provide process information, which improves the interpretability of a regression-analyzed dataset. Furthermore, the information processing device 1E can perform degradation assessment based on a trained model obtained by training using training data with little variance.

[0090] (Variation 2) The output unit 12 may output, to the display device 2, information on the process corresponding to the value determined to be deteriorated by the deterioration determination unit 17 in the dataset input to the trained model. FIG. 18 is a diagram showing an overview of the process of presenting data on deterioration candidates. In FIG. 18, the model is a learning model that simulates the time change of vibrations occurring in a railway vehicle traveling on the line indicated by route map A, and is generated by learning using dataset (1). After, for example, 10 years have passed since this model was generated, the route environment also changes, and the trained model needs to be re-created. Dataset (2) is assumed to be data obtained, for example, 10 years after the model was generated using dataset (1).

[0091] By performing learning using dataset (2), the trained model shown in the center of FIG. 18 is generated. Area B1 in route map A is an area where normal data, but outlier points outside the confidence interval, appear consecutively. Dataset (1) excludes data with variability acquired from railway vehicles traveling in area B1. In other words, the information in area B1 is information about the process in which data corresponding to the outliers extracted from the regression analysis data by the extraction unit 11 was acquired.

[0092] The deterioration determination unit 17 performs deterioration determination based on the trained model. For example, the deterioration determination unit 17 receives the data set (2) as input and outputs the result of deterioration determination using the trained model. As a result of this deterioration determination, areas B1 and B2 are determined to be deterioration candidate areas where the track is predicted to be deteriorated. The data set (2) includes data acquired from area B1 and data acquired from area B2.

[0093] The deterioration determination unit 17 outputs information indicating the area B1 and the area B2 to the output unit 12 as information on the process corresponding to the value determined to have deteriorated in the data set (2). The output unit 12 generates display information showing the area B1 and the area B2 on the route map A, and outputs the generated display information to the display device 2. Based on this display information, the display device 2 displays the route map A shown in Fig. 18. By referring to this route map A, it is possible to understand the process in which data with variability appears.

[0094] Furthermore, the degradation determination unit 17 can distinguish between a process in which data corresponding to an outlier point that is normal data but outside the confidence interval is acquired and a process in which degraded data is acquired, based on the degree of deviation from normal data. In Fig. 18, the data acquired from area B1 is not degraded data but data corresponding to an outlier, and the data acquired from area B2 is degraded data.

[0095] Therefore, since the data acquired from region B1 is not degradation data but data corresponding to an outlier, the degradation determination unit 17 may associate label information indicating that the data corresponds to an outlier with the information indicating a degradation candidate. Based on the label information, the output unit 12 generates display information showing regions B1 and B2 on the route map A along with information indicating that the data acquired from region B1 is not degradation, and outputs the generated display information to the display device 2. Based on this display information, the display device 2 displays the route map A shown in FIG. 18 . By referring to this route map A, it can be determined that region B1 is a process where data corresponding to an outlier appears. Note that the learning unit 16 may train the trained model using a dataset (2) excluding the data acquired from region B1 based on the label information.

[0096] As described above, the information processing device 1D according to the fifth embodiment includes the learning unit 16 that performs learning based on a data set excluding the values ​​extracted by the extraction unit 11, based on settings from the user or the determination result by the determination unit 14. This allows the information processing device 1D to perform learning using learning data with little variance.

[0097] The information processing device 1E according to the fifth embodiment includes a deterioration determination unit 17 that performs deterioration determination based on a trained model trained by a learning unit 16. The deterioration determination can be performed based on a trained model obtained by training using training data with little variance.

[0098] In the information processing device 1E according to the fifth embodiment, the output unit 12 outputs information on the process corresponding to the value determined to be deteriorated by the deterioration determination unit 17 in the data set input to the trained model to the display device 2. By referring to the displayed information, it is possible to grasp the process in which data with variation appears.

[0099] Sixth Embodiment In the fourth embodiment, an information processing device was described that labels information of some values ​​that are not used for learning or data corresponding to information of some values ​​that are not used for learning. In the sixth embodiment, an information processing device will be described that removes labeled data from a data set that is input to a trained model.

[0100] FIG. 19 is a block diagram showing the configuration of an information processing device 1F according to embodiment 6. In FIG. 19 , the information processing device 1F extracts information on at least some values ​​from the regression analysis results and outputs information on processes for which data corresponding to the extracted information on some values ​​has been acquired, among the processes for which a data set has been acquired. For example, the information processing device 1F acquires the regression analysis results from an external device via a network and generates process information based on the acquired information. The external device may be a regression analysis device that performs regression analysis on a preset data set. Furthermore, the information processing device 1F may acquire the regression analysis results stored in its own storage unit and generate process information based on the acquired information.

[0101] The information processing device 1F includes an extraction unit 11, an output unit 12B, a determination unit 14, a labeling unit 15, a learning unit 16, and a removal unit 18. For example, the information processing device 1F is realized by a computer. A memory included in the computer stores a program constituting an information processing application for realizing each of the functions of the extraction unit 11, the output unit 12B, the determination unit 14, the labeling unit 15, the learning unit 16, and the removal unit 18. A processor included in the computer executes the information processing application read from the memory, thereby realizing each of the functions of the extraction unit 11, the output unit 12B, the determination unit 14, the labeling unit 15, the learning unit 16, and the removal unit 18.

[0102] (Removal Unit) The removal unit 18 removes labeled data from the dataset input to the trained model. For example, the labeling unit 15 labels values ​​extracted by the extraction unit 11 and not used for learning, or data corresponding to values ​​extracted by the extraction unit 11 and not used for learning. The removal unit 18 removes data to which label information has been assigned by the labeling unit 15 from the data constituting the dataset for learning.

[0103] Next, an information processing method according to the sixth embodiment will be described. FIG. 20 is a flowchart showing the information processing method according to the sixth embodiment. The extraction unit 11 extracts at least some value information from the results of regression analysis of a dataset used for learning (step ST1E). The determination unit 14 determines whether the process information output from the extraction unit 11 complies with the set conditions (step ST2E). If the set conditions are complied with, the determination unit 14 does not use the values ​​extracted by the extraction unit 11 for learning. If the set conditions are not complied with, the determination unit 14 may use the values ​​extracted by the extraction unit 11 for learning. Based on the determination result by the determination unit 14, the labeling unit 15 labels the information on the part of values ​​not used for learning or data corresponding to the information on the part of values ​​not used for learning (step ST3E). The removal unit 18 removes labeled data from the dataset to be input to the trained model (step ST4E). The learning unit 16 performs learning using the dataset from which the part of data has been removed by the removal unit 18 (step ST5E). The output unit 12 outputs the learning results obtained by the learning unit 16 (step ST6E). For example, upon receiving the regression analysis results from the learning unit 16, the output unit 12 generates display information for displaying the regression analysis results and outputs the generated display information to the display device 2. The display device 2 displays the regression analysis results obtained by the learning unit 16 based on the display information. As in the first embodiment, the output unit 12 may cause the display device 2 to display information about processes from which data corresponding to some of the values ​​extracted by the extraction unit 11 were obtained, among the processes from which datasets were obtained. By executing the series of processes shown in FIG. 20 , the information processing device 1F can provide process information, which improves the interpretability of the regression-analyzed dataset. Furthermore, the information processing device 1F can perform learning using learning data with little variance.

[0104] As described above, the information processing device 1F according to the sixth embodiment includes the labeling unit 15 that labels values ​​extracted by the extraction unit 11 and not used for learning, or data corresponding to values ​​extracted by the extraction unit 11 and not used for learning, and the removal unit 18 that removes labeled data from a data set to be input to a trained model. This enables the information processing device 1F to perform learning using training data with little variance.

[0105] It is possible to combine the embodiments, modify any of the components of the embodiments, or omit any of the components of the embodiments.

[0106] An information processing device according to the present disclosure can be used in, for example, various systems or devices in which repetitive operations are performed.

[0107] 1, 1A, 1B, 1C, 1D, 1E, 1F Information processing device, 2 Display device, 3 Input device, 4 Learning device, 11, 11A Extraction unit, 12, 12A, 12B Output unit, 13 Selection unit, 14 Determination unit, 15 Labeling unit, 16 Learning unit, 17 Degradation determination unit, 18 Removal unit, 100 Input interface, 101 Output interface, 102 Processor, 103 Memory.

Claims

1. An information processing device comprising: an extraction unit that extracts information on at least some values ​​from the results of regression analysis of a dataset used for learning; and an output unit that outputs information on a process from which data corresponding to the information on the some values ​​extracted by the extraction unit was obtained, among the processes from which the dataset was obtained.

2. The information processing device according to claim 1, characterized in that the extraction unit extracts information on the part of values ​​determined based on predicted values ​​of the dependent variable for the explanatory variables.

3. The information processing device according to claim 2, characterized in that the extraction unit extracts information on the part of values ​​that are outside a confidence interval that includes a predicted value of the dependent variable for the explanatory variables.

4. The information processing device according to claim 1, characterized in that the extraction unit extracts information on the part of values ​​determined by a user from the results of regression analysis of the data set.

5. The information processing device according to claim 4, characterized in that the extraction unit extracts information on the part of values ​​that indicate values ​​outside a confidence interval determined by a user, which includes the predicted value of the dependent variable for the explanatory variable.

6. The information processing device according to any one of claims 1 to 5, wherein the data set is obtained by repeatedly performing a process.

7. An information processing device according to any one of claims 1 to 6, characterized in that the output unit outputs to a display device information on a process in which data corresponding to the information on the part of values ​​extracted by the extraction unit was obtained.

8. The information processing device according to claim 7, characterized in that the output unit outputs to a display device information indicating the relationship between information on the partial value and information on the process in which data corresponding to the information on the partial value extracted by the extraction unit was obtained.

9. An information processing device according to any one of claims 1 to 8, further comprising a selection unit that accepts a selection as to whether or not information on the partial value extracted by the extraction unit is to be used for learning.

10. An information processing device according to any one of claims 1 to 8, characterized in that it comprises a judgment unit that judges whether the information of the process from which data corresponding to the information of the part of the values ​​extracted by the extraction unit was obtained complies with set conditions, and if it complies with the set conditions, the value extracted by the extraction unit is not used for learning.

11. An information processing device as described in claim 10, characterized in that it is provided with a labeling unit that labels the information on the part of values ​​extracted by the extraction unit and not used for learning, or data corresponding to the information on the part of values ​​extracted by the extraction unit and not used for learning.

12. The information processing device according to claim 10, wherein values ​​extracted by said extraction unit are learned using a different learning method from values ​​not extracted.

13. An information processing device according to claim 10, further comprising a learning unit that performs learning based on the data set excluding the values ​​extracted by the extraction unit, based on settings from a user or the judgment result by the judgment unit.

14. The information processing device according to claim 13, further comprising a deterioration determination unit that determines deterioration based on a trained model trained by said learning unit.

15. The information processing device according to claim 14, characterized in that the output unit outputs to a display device information on a process corresponding to a value determined to be deteriorated by the deterioration determination unit in the data set input to the trained model.

16. The information processing device according to claim 14, further comprising: a labeling unit that labels values ​​extracted by the extraction unit and not used for learning, or data corresponding to values ​​extracted by the extraction unit and not used for learning; and a removal unit that removes labeled data from the dataset input to the trained model.

17. An information processing device according to any one of claims 1 to 16, characterized in that the extraction unit sets consecutive intervals based on the size of the intervals between each value in the process among the values ​​obtained by regression analysis of the data set, and extracts information on values ​​included in the set intervals.

18. An information processing method executed by an information processing device, comprising: a step in which an extraction unit extracts information on at least some values ​​from the results of regression analysis of a dataset used for learning; and a step in which an output unit outputs information on a process in which data corresponding to the values ​​extracted by the extraction unit was obtained, among the processes in which the dataset was obtained.

19. A program for causing a computer to execute the steps of: an extraction unit extracting information on at least some values ​​from the results of regression analysis of a dataset used for learning; and an output unit outputting information on a process, among the processes in which the dataset was obtained, in which data corresponding to the values ​​extracted by the extraction unit was obtained.

Images