JP7500980B2 - Information processing device, information processing method, and information processing program - Google Patents

Description

The present invention relates to an information processing device, an information processing method, and an information processing program.

In recent years, there has been an increase in cases where abnormalities in various devices are detected by classifying data using machine learning (for example, Patent Document 1).

JP 2019-056983 A

However, when using a learning model based on deep learning, for example, it is generally difficult to identify the cause of an anomaly when one is detected.

The object of the present invention is to provide an information processing device capable of estimating features that are factors in classification.

One invention for achieving the above object is an information processing device including a clustering unit that groups a plurality of first data including a plurality of variables into a plurality of clusters, a generating unit that assigns identification information for identifying the data included in each of the plurality of clusters to each of the plurality of first data and generates a plurality of second data, a model constructing unit that constructs, based on the plurality of second data, a first learning model for classifying input data into one of a plurality of patterns based on the data included in each of the plurality of clusters, and a calculating unit that calculates, based on the plurality of second data and the first learning model, influence information indicating the influence of each of the plurality of variables in each of the plurality of patterns. Other features of the present invention will be made clear by the description in this specification.

The present invention provides an information processing device capable of estimating features that are factors in classification.

FIG. 10 is a diagram showing the configuration of an information processing system. FIG. 2 is a diagram illustrating an example of a hardware configuration of an information processing device 20. FIG. 4 is a diagram showing an example of first data 41. FIG. 4 is a diagram for explaining first data 41. FIG. 2 is a diagram illustrating an example of functional blocks realized in the information processing device 20. 1 is a scatter plot showing an example of first data 41 grouped into a plurality of clusters. FIG. 4 is a diagram showing an example of second data 42. A diagram for explaining the first learning model M1. FIG. 2 is a diagram showing an example of impact information 70. A figure for explaining the first data 41 and the second learning model M2. 4 is a flowchart showing an example of processing executed by the information processing device 20. FIG. 2 is a diagram illustrating an example of a hardware configuration of a diagnostic device 21. FIG. 2 is a diagram showing an example of a functional block realized in a diagnostic device 21. 4 is a flowchart showing an example of a process executed by a diagnostic device 21.

== ...
<<<<Configuration of Information Processing System 10>>>
1 is a diagram showing the configuration of an information processing system 10 according to an embodiment of the present invention. The information processing system 10 is a system for estimating features that are factors in causing an abnormality in a predetermined device, and includes an information processing device 20 and a diagnostic device 21. In this embodiment, a description will be given using a showcase 300 installed in a commercial facility as an example of the predetermined device.

The showcase 300 is, for example, a case for cooling and storing food and the like. The showcase 300 is equipped with a number of sensors 310 that monitor the state of the showcase 300. For convenience, the multiple sensors 310 are depicted as a single block in FIG. 1.

The information processing system 10 then diagnoses whether the showcase 300 is normal or abnormal based on the data output from each of the multiple sensors 310. Furthermore, in the diagnostic process, the information processing system 10 outputs the degree of influence (described below) of the data output from each of the multiple sensors 310.

In this case, examples of "abnormalities" include cases where the temperature inside showcase 300 cannot be maintained at a specified temperature, or where showcase 300 makes an abnormal noise, or where other behavior that differs from normal is observed.

Furthermore, hereinafter, data when the showcase 300 is operating normally will be referred to as "normal data" or "normal data", and data when the showcase 300 is operating abnormally will be referred to as "abnormal data" or "abnormal data".

The information processing device 20 uses machine learning to construct a model for determining whether or not there is an abnormality in the showcase 300 based on the data output by the multiple sensors 310.

The diagnostic device 21 diagnoses whether or not there is an abnormality in the showcase 300 during operation based on data output from the multiple sensors 310 and a model constructed by the information processing device 20. Furthermore, the diagnostic device 21 outputs the degree of influence of the data output from each of the multiple sensors 310 in the diagnostic process. The information processing device 20 and the diagnostic device 21 are connected via a network 25.

<<<<About the information processing device 20>>>
==Configuration of Information Processing Device 20==
2 is a diagram showing an example of a hardware configuration of the information processing device 20. The information processing device 20 is a computer including a CPU (Central Processing Unit) 30, a memory 31, a storage device 32, an input device 33, a display device 34, and a communication device 35.

The CPU 30 executes programs stored in the memory 31 and the storage device 32 to realize various functions in the information processing device 20.

Memory 31 is, for example, a RAM (Random Access Memory) and is used as a temporary storage area for programs, data, etc.

The storage device 32 is a non-volatile storage device that stores various data such as the control program 40 executed or processed by the CPU 30 and the first data 41 (described below).

The control program 40 is a program for implementing various functions of the information processing device 20, and includes, for example, an OS (Operating System), etc.

The multiple first data 41 are data used when constructing the first learning model M1 or the second learning model M2 (described later). FIG. 3 is a diagram showing the first data 41 in this embodiment. Each of the multiple first data 41 includes data on multiple variables (four variables in this embodiment) x1 to x4. In addition, the multiple first data 41 are data having a predetermined attribute. In this embodiment, the predetermined attribute is information indicating whether or not a predetermined device is operating normally.

Here, "variable x1" and "variable x2" respectively refer to the temperatures indicated by, for example, a first temperature sensor and a second temperature sensor attached to a predetermined location of the showcase 300. "Data of variable x1" and "data of variable x2" are values output from the first temperature sensor and the second temperature sensor, respectively.

"Variable x3" means, for example, the pressure indicated by a pressure sensor that measures the pressure of the compressor inside showcase 300. "Data of variable x3" is the value output from the pressure sensor. Furthermore, "Variable x4" means, for example, the flow rate indicated by a flow meter that measures the flow rate of the refrigerant in the compressor. "Data of variable x4" is the value output from the flow meter.

The "attribute" is data that indicates whether the data of variables x1 to x4 is "normal data" or "abnormal data." In this embodiment, if the data of variables x1 to x4 is "normal data," a "0" is assigned, and if the data of variables x1 to x4 is "abnormal data," a "1" is assigned.

In this embodiment, the first data 41 includes i pieces of "normal data" and j pieces of "abnormal data", and is stored in advance in the storage device 32. The first data of the first data 41 is, for example, data of variables x1 to x4 obtained at time t1 when the showcase 300 is normal.

Figure 4 is a scatter plot showing the first data 41 shown in Figure 3. Note that each of the multiple first data 41 includes data and attributes for four variables x1 to x4, but for convenience, Figure 4 illustrates the data as being for two variables, x1 and x2.

The input device 33 is a device that accepts commands and data input by the user, and includes input interfaces such as a keyboard and a touch sensor that detects the touch position on a touch panel display.

The display device 34 is, for example, a device such as a display, and the communication device 35 exchanges various programs and data with the diagnostic device 21 and other computers via the network 25.

Example of Functional Blocks
5 is a diagram showing an example of functional blocks realized in the information processing device 20. When the CPU 30 of the information processing device 20 executes the control program 40, a clustering unit 50, a generating unit 51, a model constructing unit 52, a calculating unit 53, and a first acquiring unit 54 are realized in the information processing device 20.

The clustering unit 50 groups the multiple first data 41 into multiple clusters. Note that "grouping" here means dividing data into one or more groups.

Specifically, the clustering unit 50 first acquires the first data 41 stored in the storage device 32. Then, the clustering unit 50 groups each of the "normal data" and "abnormal data" into multiple clusters.

As an algorithm for grouping into multiple clusters, an algorithm that pre-assumes the number of clusters, such as k-means, may be used, or an algorithm that does not pre-assume the number of clusters, such as x-means, may be used. Other algorithms may be used as long as they are capable of grouping into multiple clusters. In this embodiment, the clustering unit 50 processes the multiple first data 41 using x-means.

Figure 6 is a diagram for explaining a state in which "normal data" and "abnormal data" are each classified into multiple clusters. Here, "normal data" is grouped into two clusters, C1 and C2, and "abnormal data" is grouped into two clusters, C3 and C4.

If the clustering unit 50 groups all of the first data 41 into multiple clusters without separating the "normal data" and "abnormal data" of the first data 41, one cluster may contain both "normal data" and "abnormal data." In such a case, if the data contained in the cluster is used as learning data, it becomes unclear whether the learning data is data indicating "normal data" or data indicating "abnormal data."

In this embodiment, the first data 41 is separated into "normal data" and "abnormal data," and each of the data is grouped into multiple clusters. Therefore, in this embodiment, the accuracy of the learning data can be improved.

The clustering unit 50 does not have to group both the "normal data" and the "abnormal data" of the first data 41 into multiple clusters. For example, only the "abnormal data" may be grouped into multiple clusters.

The generation unit 51 assigns identification information for identifying the data contained in each of the multiple clusters to each of the multiple first data 41, thereby generating multiple second data 42. FIG. 7 is a diagram showing the second data 42 of this embodiment. In this example, among the first data 41, the first data 41 grouped into cluster C1 is assigned identification information "1", and the first data 41 grouped into cluster C2 is assigned identification information "2".

Furthermore, the first data 41 grouped into cluster C3 is given the identification information "3", and the first data 41 grouped into cluster C4 is given the identification information "4", thereby generating second data 42.

The model construction unit 52 constructs a first learning model M1 based on the multiple second data 42. The first learning model M1 is a learning model for classifying input data into one of multiple patterns. The multiple patterns are based on the data contained in each of the multiple clusters.

In this embodiment, the multiple patterns consist of Pattern 1 to Pattern 4. Pattern 1 is based on the data contained in Cluster C1, Pattern 2 is based on the data contained in Cluster C2, Pattern 3 is based on the data contained in Cluster C3, and Pattern 4 is based on the data contained in Cluster C4.

In this embodiment, when the first learning model M1 is learned, the coefficients of the function of the first learning model M1 are adjusted. Note that the first learning model M1 is constructed based on, for example, a support vector machine (SVM) technique, and the function of the first learning model M1 is expressed as, for example, y = f1 (x1, x2, x3, x4).

In FIG. 8, an example of a function f1 representing a first learning model M1 that classifies multiple second data 42 into one of four patterns is illustrated on the x1-x2 plane.

The model construction unit 52 further constructs a second learning model M2 based on the data acquired by the first acquisition unit 54 described below. Like the first learning model M1, the second learning model M2 is a learning model for classifying input data as to which of a plurality of patterns it belongs to. In other words, the second learning model M2 of this embodiment also classifies input data as to which of patterns 1 to 4 it belongs to.

The second learning model M2 is a learning model that is output to the diagnostic device 21. Details of how to construct the second learning model M2 will be described later.

The calculation unit 53 calculates the influence information 70. The influence information 70 is information indicating the influence of each of the multiple variables in each of the multiple patterns.

Here, the influence level refers to the influence level of each of the multiple variables contained in the classified second data 42 on the result of classification based on the first learning model M1. The influence level may be expressed as a value indicating the influence level in a predetermined range, for example, 0 to 100%. In addition, the influence level may be expressed as a binary value such as "strong" or "weak" with a predetermined value set as a threshold for the value indicating the influence level, or may be expressed as a multi-value value of three or more.

The information indicating the degree of influence is information including at least a portion of the multiple variables and the respective degrees of influence of the multiple variables. In other words, the information indicating the degree of influence may include all of the multiple variables and the respective degrees of influence of the multiple variables. Furthermore, the information indicating the degree of influence may be only the variable corresponding to the highest value indicating the degree of influence.

Figure 9 is a diagram showing the impact information 70 of this embodiment. The impact information 70 of this embodiment is information that includes values indicating the impact of each of multiple variables in each of multiple patterns. In this example, the values indicating the impact are shown in the range of 0 to 100%. Furthermore, in each pattern, the sum of the impacts of each of the variables x1 to x4 is normalized to be 100%.

Hereinafter, the "value indicating the degree of influence" will be simply referred to as the "degree of influence." In this embodiment, the degree of influence of a variable xl (l=1 to 4) in a pattern k (k=1 to 4) is defined as the degree of influence A _kl .

The influence A _kl in this embodiment is the influence output for the second data 42 classified into pattern k by the first learning model M1. That is, the influence A _kl in this embodiment is an influence that is approximated so as to be constant regardless of the data of the variable xl (l = 1 to 4) as long as the data of the input variable xl (l = 1 to 4) is classified into pattern k. A method for calculating the influence A _kl will be described in detail later.

For example, pattern 3 shown in the impact information 70 in FIG. 9 is based on the second data 42 included in cluster C3 (FIG. 6). The second data 42 included in cluster C3 is data that is assigned the attribute "1" indicating "abnormal data" and the identification information "3" (FIG. 7).

FIG. 9 shows that in pattern 3, the influence of variable x1 (A ₃₁ ) is 10%, the influence of variable x2 (A ₃₂ ) is 2%, the influence of variable x3 (A ₃₃ ) is 80%, and the influence of variable x4 (A ₃₄ ) is 8%.

In this example, when the data of the input variable xl (l = 1 to 4) is classified into pattern 3, the influence of the variable x3 is dominant. This means that in the classification using the first learning model M1, the data of the variable x3 is considered to be the most important result. This also means that it can be estimated that the output value of the pressure sensor representing the variable x3 is abnormal.

The following describes a method for calculating a value indicating the impact information 70 in this embodiment. The value indicating the impact information 70 is calculated based on a plurality of second data 42 and the first learning model M1.

First, the calculation unit 53 calculates the degree of influence for each of the multiple second data 42 using an algorithm that can be used as a method for interpreting the machine learning model. As the algorithm, for example, an algorithm such as LIME (local interpretable model-agnostic explanations) or SHAP (SHapley Additive exPlanations) can be used. In this embodiment, the degree of influence is calculated using LIME.

Specifically, using LIME, the influence a kl (l = 1 to 4) of the variable xl (l = 1 to 4) is calculated for one second data 42 to which identification information "k" (k = 1 to 4) is assigned among the multiple second data 42. At this time, a local linear approximation model is generated based on the first learning model M1 around the one second data 42, and the _weight w _kl (l = 1 to 4) of the variable xl (l = 1 to 4) in the linear approximation model is calculated.

Then, the calculated weights w _kl (k=1 to 4, l=1 to 4) are normalized so that the sum for l is 100, and the result is defined as the influence a _kl (k=1 to 4, l=1 to 4) of the variable xl (l=1 to 4) for one second data 42. The above calculation process is performed for all second data 42.

Next, the influence a _kl (l=1 to 4) is added up for all the second data 42 to which identification information "k" (k=1 to 4) is assigned among the multiple second data 42, and the sum is set as the weight W _kl (l=1 to 4). The calculated weights W _kl (l=1 to 4) are then normalized so that the sum for l is 100, and the above-mentioned influence A kl (l=1 to 4) is set as the above-mentioned influence A _kl (l=1 to 4).

The first acquisition unit 54 acquires a portion of the second data 42, including at least one piece of data contained in each of the multiple clusters. As described above, the second data 42 acquired by the first acquisition unit 54 from the multiple second data 42 is used by the model construction unit 52 to construct the second data 42.

Figure 10 is a diagram showing the second data 42 acquired by the first acquisition unit 54. In Figure 10, black circles indicate the second data 42 acquired by the first acquisition unit 54 among the multiple pieces of second data 42. On the other hand, white circles indicate the second data 42 other than the second data 42 acquired by the first acquisition unit 54 among the multiple pieces of second data 42.

The first acquisition unit 54 acquires at least the data closest to the center of gravity of each of the plurality of clusters among the plurality of second data 42. Note that the "center of gravity" of a cluster here is the average value of the plurality of second data 42 included in the cluster.

In the example of FIG. 10, the first acquisition unit 54 acquires the second data 42 (black circles) that exist within a predetermined distance from the center of gravity of each of the multiple clusters. The distance here is not particularly limited, but for example, the Euclidean distance, Manhattan distance, Chebyshev distance, etc. can be used.

The second learning model M2 is constructed based on a method similar to that of the first learning model M1 described above, and the function of the second learning model M2 is expressed as, for example, y = f2 (x1, x2, x3, x4).

In other words, the second learning model M2 is constructed based on representative second data 42 that exists near the center of gravity of each of the multiple clusters, among the multiple second data 42. As a result, the model construction unit 52 trains the second learning model M2 using a small amount of data while maintaining information about the distribution of the second data 42. Note that the "distribution of data" refers to, for example, the distribution in vector space when each of the data x1 to x4 is treated as a vector.

The second learning model M2 constructed in this way can be said to be a learning model that approximately represents the first learning model M1. Furthermore, since the second learning model M2 is constructed using a smaller amount of data than the first learning model M1, the function f2 of the second learning model M2 has a smaller amount of data than the function f1 of the first learning model M1.

The second learning model M2 is used by the classification unit 81 in the diagnostic device 21 described below in a process of classifying diagnostic data into one of a number of patterns using an SVM. In this classification process, by using the second learning model M2, which has a smaller amount of data than the first learning model M1, the processing speed can be increased compared to when the first learning model M1 is used.

<<Information Processing S10>>
An example of the process executed by each functional block will be described below with reference to Fig. 11 etc. Fig. 11 is a flowchart showing an example of the process executed by the information processing device 20.

First, the clustering unit 50 acquires the first data 41 (FIGS. 3 and 4) stored in the storage device 32 (S11).

Next, the clustering unit 50 groups the first data 41 into multiple clusters (S12, FIG. 6). At this time, the clustering unit 50 separates the "normal data" and "abnormal data" of the first data 41 and groups each of them into multiple clusters.

Next, the generation unit 51 assigns identification information for identifying the data contained in each of the multiple clusters to each of the multiple first data 41, thereby generating multiple second data 42 (S13, Figure 7).

Next, the model construction unit 52 constructs a first learning model M1 based on the multiple second data 42 generated in S13 (S14, Figure 8).

Next, the calculation unit 53 calculates the impact information 70 based on the multiple second data 42 generated in S13 and the first learning model M1 constructed in S14 (S15, Figure 9).

Next, the first acquisition unit 54 acquires a portion of the second data 42 that includes at least one piece of data contained in each of the multiple clusters (S16, FIG. 10).

Next, the model construction unit 52 constructs a second learning model M2 based on a portion of the multiple second data 42 acquired in S16 (S17, FIG. 10).

Next, the information processing device 20 outputs the impact information 70 calculated in S15 and the second learning model M2 constructed in S17 to the diagnostic device 21.

<<<About the diagnosis device 21>>>
==Configuration of Diagnosis Device 21==
12 is a diagram showing an example of a hardware configuration of the diagnostic device 21. The diagnostic device 21 is a computer including a CPU 60, a memory 61, a storage device 62, an input device 63, a display device 64, and a communication device 65. Note that the hardware configuration of the diagnostic device 21 is similar to that of the information processing device 20, and therefore a detailed description thereof will be omitted here.

The storage device 62 stores the second learning model M2, the impact information 70, and the diagnostic program 71. The second learning model M2 is a model constructed by the information processing device 20.

The diagnostic program 71, like the control program 40, is a general term for the programs for realizing the various functions of the diagnostic device 21.

==Function Blocks==
12 is a diagram showing an example of functional blocks realized in the diagnostic device 21. When the CPU 60 of the diagnostic device 21 executes the diagnostic program 71, a second acquisition unit 80, a classification unit 81, an output unit 82, and a storage unit 83 are realized in the diagnostic device 21.

The second acquisition unit 80 acquires diagnostic data that includes multiple variables and is to be diagnosed. Specifically, the second acquisition unit 80 acquires data of variables x1 to x4 output from the sensor 310 of the showcase 300 during operation at predetermined time intervals (e.g., every 30 seconds). The data of variables x1 to x4 output from the sensor 310 corresponds to "diagnostic data."

The classification unit 81 classifies the diagnostic data into one of a plurality of patterns based on the second learning model M2 constructed by the model construction unit 52 of the information processing device 20. In this embodiment, the plurality of patterns consists of pattern 1 to pattern 4, as described above.

The output unit 82 outputs the influence information 70 ( FIG. 9 ) of the pattern into which the diagnostic data has been classified. For example, when the diagnostic data acquired at a certain time is classified into pattern 4 by the classification unit 81, in this embodiment, the output unit 82 outputs that the influence (A ₄₁ ) of the variable x1 is 1%, the influence (A ₄₂ ) of the variable x2 is 4%, the influence (A ₄₃ ) of the variable x3 is 0%, and the influence (A ₄₄ ) of the variable x4 is 95%. At this time, the output unit 82 may output to the display device 64 or may output by audio information.

Then, the worker confirms the output from the output unit 82 and realizes that an abnormality has been detected in the showcase 300. In this case, the worker realizes that the main cause of the abnormality is the flow rate of the refrigerant in the compressor, which is represented by the variable x4.

The storage unit 83 stores the impact information 70 output from the information processing device 20 to the diagnostic device 21. This allows previously generated impact information 70 to be used in the diagnostic process. This reduces the number of times the information processing device 20 generates the second learning model M2 and the impact information 70. This also eliminates the need to calculate the impact for each piece of diagnostic data using LIME or the like, reducing the calculation load in the diagnostic process.

In this embodiment, the classification unit 81 classifies the diagnostic data into one of a plurality of patterns based on the second learning model M2, but this is not limited to the above. The classification unit 81 may classify the diagnostic data based on the first learning model M1 instead of the second learning model M2.

In this case, the information processing device 20 does not need to construct the second learning model M2. In addition, in this case, the information processing device 20 outputs the first learning model M1 to the diagnostic device 21 instead of the second learning model M2.

<<Diagnosis Process S20>>
An example of the process executed by each functional block will be described below with reference to Fig. 14 etc. Fig. 14 is a flow chart showing an example of the process executed by the diagnostic device 21.

First, as shown in FIG. 14, the second acquisition unit 80 acquires diagnostic data that includes multiple variables and is to be diagnosed (S21). Next, the classification unit 81 classifies the diagnostic data into one of multiple patterns based on the second learning model M2 (S22). Next, the output unit 82 outputs impact information 70 of the classified patterns of the diagnostic data (S23).

====Summary====
The above describes the information processing system 10 according to the present embodiment. According to this embodiment, it is possible to provide an information processing system capable of estimating features that are factors in classification.

Furthermore, according to this embodiment, when the clustering unit 50 groups the multiple first data 41 into multiple clusters, it performs x-means processing. Therefore, it is not necessary to assume the number of clusters. This eliminates preconceptions about the number of features when estimating the features that are factors in classification, allowing for accurate estimation.

Furthermore, according to this embodiment, the multiple first data 41 are data having a predetermined attribute. Therefore, when the clustering unit 50 groups the multiple first data 41 into multiple clusters, the multiple first data 41 can be grouped by attribute. This can improve the accuracy of the multiple first data 41 as learning data.

Furthermore, according to this embodiment, the specified attribute is information indicating whether or not the specified device is operating normally. In other words, the clustering unit 50 can group the multiple first data 41 into multiple clusters while separating the "normal data" and "abnormal data". Therefore, "normal data" and "abnormal data" are not mixed in one cluster, and the accuracy of the multiple first data 41 as learning data can be improved.

Furthermore, according to this embodiment, the influence information 70 is information that includes values indicating the influence of each of a plurality of variables. This makes it possible to quantitatively estimate the characteristics that are the factors behind the classification.

Furthermore, according to this embodiment, the first acquisition unit 54 acquires a portion of the second data 42, including at least one piece of data contained in each of the multiple clusters. This makes it possible to train the second learning model M2 using a smaller amount of data while maintaining information about the distribution of the second data 42. This makes it possible to reduce the training time of the second learning model M2.

Furthermore, according to this embodiment, the first acquisition unit 54 acquires at least the data closest to the center of gravity of each of the plurality of clusters among the plurality of second data 42. This makes it possible to efficiently reduce the amount of data used to train the second learning model M2 while maintaining information about the distribution of the second data 42.

Furthermore, according to this embodiment, in the diagnostic process, for each piece of diagnostic data output from a specific device, the characteristics that are factors in the classification can be estimated by referring to the influence information 70 based on the results of classification by the second learning model M2.

In conventional anomaly detection methods, the degree of impact is calculated for each piece of diagnostic data output from the device during diagnostic processing using existing interpretation methods for machine learning models such as LIME and SHAP. This poses the problem of a huge computational load during diagnostic processing.

However, according to the diagnostic process of this embodiment, there is no need to calculate the degree of influence for each piece of diagnostic data. This reduces the calculation load in the diagnostic process.

Furthermore, according to this embodiment, the diagnostic device 21 includes a storage unit 83 that stores the impact information 70. This allows previously generated impact information 70 to be used in the diagnostic process. This reduces the number of times the information processing device 20 generates the second learning model M2 and the impact information 70.

The above embodiment is intended to facilitate understanding of the present invention, and is not intended to limit the present invention. Furthermore, the present invention may be modified or improved without departing from the spirit of the present invention, and it goes without saying that the present invention includes equivalents.

10: Information processing system 20: Information processing device 21: Diagnostic device 25: Network 30: CPU
31: Memory 32: Storage device 33: Input device 34: Display device 35: Communication device 40: Control program 41: First data 42: Second data 50: Clustering unit 51: Generation unit 52: Model construction unit 53: Calculation unit 54: First acquisition unit 60: CPU
61: Memory 62: Storage device 63: Input device 64: Display device 65: Communication device 70: Impact information 71: Diagnostic program 80: Second acquisition unit 81: Classification unit 82: Output unit 83: Storage unit 300: Showcase 310: Sensor M1: First learning model M2: Second learning model

Claims (11)

A clustering unit that groups a plurality of first data including a plurality of variables into a plurality of clusters;
a generating unit that generates a plurality of second data by adding, to each of the plurality of first data, identification information that identifies data included in each of the plurality of clusters for each of the plurality of clusters;
a model construction unit that constructs a first learning model based on the plurality of second data for classifying input data into one of a plurality of patterns based on data included in each of the plurality of clusters;
A calculation unit that calculates influence information indicating an influence of each of the plurality of variables , which are characteristics that are factors in classifying the plurality of patterns, based on the plurality of second data and the first learning model;
An information processing device comprising: 2. The information processing device according to claim 1,
The clustering unit performs x-means processing on the plurality of first data;
An information processing device comprising: 3. The information processing device according to claim 1,
the plurality of first data are data having a predetermined attribute;
An information processing device comprising: 4. The information processing device according to claim 3,
the predetermined attribute is information indicating whether a predetermined device is operating normally;
An information processing device comprising: The information processing device according to any one of claims 1 to 4,
the influence information includes a value indicating an influence of each of the plurality of variables;
An information processing device comprising: The information processing device according to any one of claims 1 to 5,
a first acquisition unit that acquires a portion of the second data including at least one piece of data included in each of the plurality of clusters,
The model construction unit includes:
constructing a second learning model for classifying input data into which of the plurality of patterns the data belongs, based on the data acquired by the first acquisition unit;
An information processing device comprising: 7. The information processing device according to claim 6,
The first acquisition unit is
acquiring at least data closest to a center of gravity of each of the plurality of clusters from among the plurality of second data;
An information processing device comprising: An information processing device according to claim 6 or 7;
A second acquisition unit that acquires diagnostic data that includes the plurality of variables and is to be diagnosed;
a classification unit that classifies the diagnostic data into one of the plurality of patterns based on the second learning model;
an output unit that outputs the impact information of a pattern in which the diagnostic data is classified;
and a diagnostic device including the above. 9. The information processing system according to claim 8,
the diagnostic device includes a storage unit that stores the impact information;
An information processing system comprising: Grouping a plurality of first data including a plurality of variables into a plurality of clusters;
a step of generating a plurality of second data by assigning identification information to each of the plurality of first data, the identification information being used to identify data included in each of the plurality of clusters for each of the plurality of clusters;
constructing a first learning model based on the plurality of second data for classifying input data into one of a plurality of patterns based on data included in each of the plurality of clusters;
Calculating influence information indicating an influence of each of the plurality of variables , which are characteristics that are factors in classifying the plurality of patterns, based on the plurality of second data and the first learning model;
An information processing method comprising: On the computer,
Grouping a plurality of first data including a plurality of variables into a plurality of clusters;
assigning identification information to each of the plurality of first data, the identification information identifying the data included in each of the plurality of clusters for each of the plurality of clusters, to generate a plurality of second data;
constructing a first learning model based on the plurality of second data, the first learning model being used to classify input data into one of a plurality of patterns based on data included in each of the plurality of clusters;
calculating influence information indicating an influence of each of the plurality of variables , which are characteristics that are factors in classifying the plurality of patterns, based on the plurality of second data and the first learning model;
23. An information processing program comprising:

Images