Information processing device, information processing method, and program

Abstract

The information processing device (100) is an information processing device that modifies clustering results, which are the result of clustering time-series data, and comprises a change command receiving unit (110) that receives a command to change the number of clusters, and a cluster number changing unit (120) that, upon receiving a command to change the number of clusters, determines and executes the merging or splitting of clusters in the clustering results based on the clustering results and the command to change the number of clusters, and outputs the clustering results after the number of clusters has been changed.

Description

Technical Field

[0001] The disclosed technology relates to an information processing technology for changing clustering results, which are the results of clustering data.

Background Art

[0002] Clustering is an analysis method for classifying data into units such as clusters based on their respective features. In clustering technology, for example, an effective feature representation is extracted from time-series data, and unsupervised time-series segmentation is realized by clustering according to the automatically determined number of clusters K using DPGMM (Dirichlet Process Gaussian Mixture Model). On the other hand, the clusters obtained by such clustering may not always be appropriate for the user. In contrast, Patent Document 1 discloses a work analysis technology for performing clustering only on some of the clusters after clustering. Specifically, it is described that "the user can perform clustering (additional processing) with a changed granularity only for the additional processing target by selecting and operating any one of the clusters displayed in the clustering result area R22 on the work analysis screen G2 as the additional processing target."

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, the technology described in Patent Document 1 has a problem in that, for example, the classification criteria for clusters subjected to additional processing may differ significantly from the classification criteria for clusters not subjected to additional processing, and the consistency of clustering may not be maintained.

[0005] This disclosure aims to solve the above-mentioned problems by enabling the consistency of clustering before and after a change in the number of clusters in the clustering results. [Means for solving the problem]

[0006] The information processing device disclosed herein is An information processing device that modifies clustering results, which are the result of clustering time-series data, A change command receiving unit that receives a command to change the number of clusters, which indicates the commanded number of clusters, When the command to change the number of clusters is received, the system will decide whether to merge or split the clusters in the clustering result based on the clustering result and the command to change the number of clusters. Based on the characteristics of each cluster in the clustering results, the clusters to be merged or split are determined. The cluster number change section outputs the clustering results after changing the number of clusters, It is something that is provided. [Effects of the Invention]

[0007] According to this disclosure, when changing the number of clusters in the clustering results, it is possible to maintain the consistency of the clustering before and after the change. [Brief explanation of the drawing]

[0008] [Figure 1] Figure 1 shows an example of the configuration of an information processing device and peripheral devices related to the information processing device according to Embodiment 1 of the present disclosure. [Figure 2] Figure 2 shows an example configuration of an information processing device and an information processing system including the information processing device according to Embodiment 1 of this disclosure. [Figure 3]Figure 3 shows an example of the configuration of the cluster number changing unit in the information processing device according to Embodiment 1 of this disclosure. [Figure 4] Figure 4 is a flowchart showing an example of processing in an information processing device according to Embodiment 1 of this disclosure. [Figure 5] Figure 5 is a flowchart showing a detailed example of the time-series division process in the processing of the information processing device according to Embodiment 1 of this disclosure. [Figure 6] Figure 6 is a flowchart showing a detailed example of clustering processing in the information processing apparatus according to Embodiment 1 of this disclosure. [Figure 7] Figure 7 is a flowchart showing a detailed example of the process for changing the number of clusters in the information processing device according to Embodiment 1 of this disclosure. [Figure 8] Figure 8 is a flowchart showing a more detailed example of the process shown in Figure 7. [Figure 9] Figure 9 is a flowchart showing a more detailed example of the process for changing the number of clusters in the information processing device according to Embodiment 1 of this disclosure. [Figure 10] Figure 10 shows an illustration of unsupervised time series segmentation and clustering. [Figure 11] Figure 11 shows an image illustrating how to change the number of clusters in the information processing device of this disclosure. [Figure 12] Figure 12 shows an image of what happens when you try to change the number of clusters in a configuration that does not employ the information processing device of this disclosure. [Figure 13] Figure 13 shows an example of a user interface (image) for an output destination device output from the information processing device of this disclosure (an example before changing the number of clusters). [Figure 14] Figure 14 shows an example of a user interface (image) for an output destination device output from the information processing device of this disclosure (an example after changing the number of clusters). [Figure 15]FIG. 15 is a diagram showing a configuration example of an information processing apparatus and an information processing system including the information processing apparatus according to Embodiment 2 of the present disclosure. [Figure 16] FIG. 16A is a diagram showing an example of the configuration of a segment number changing unit in the information processing apparatus according to Embodiment 2 of the present disclosure, and FIG. 16B is a diagram showing an example of the configuration of a cluster number changing unit in the information processing apparatus according to Embodiment 2 of the present disclosure. [Figure 17] FIG. 17 is a flowchart showing an example of the processing of the information processing apparatus according to Embodiment 2 of the present disclosure. [Figure 18] FIG. 18 is a flowchart showing a more detailed example of the process of changing the number of segments in the process shown in FIG. 17. [Figure 19] FIG. 19 is a flowchart showing a more detailed example of the process of changing the number of segments in the information processing apparatus according to Embodiment 2 of the present disclosure. [Figure 20] FIG. 20 is a flowchart showing a more detailed example of the process of changing the number of clusters in the process shown in FIG. 17. [Figure 21] FIG. 21 is a flowchart showing a more detailed example of the process of changing the number of clusters in the information processing apparatus according to Embodiment 2 of the present disclosure. [Figure 22] FIG. 22 is a diagram showing a first configuration example of an information processing apparatus and an information processing system including the information processing apparatus according to Embodiment 3 of the present disclosure. [Figure 23] FIG. 23 is a diagram showing a second configuration example of an information processing apparatus and an information processing system including the information processing apparatus according to Embodiment 3 of the present disclosure. [Figure 24] FIG. 24 is a flowchart showing an example of the process of changing the number of segments in the information processing apparatus according to Embodiment 3 of the present disclosure. [Figure 25] FIG. 25 is a flowchart showing an example of the process of changing the number of clusters in the information processing apparatus according to Embodiment 3 of the present disclosure. [Figure 26]Figure 26 is a flowchart showing a detailed example of the process for changing the number of segments in the information processing device according to Embodiment 3 of this disclosure. [Figure 27] Figure 27 is a flowchart showing a detailed example of the process for changing the number of clusters in the information processing device according to Embodiment 3 of this disclosure. [Figure 28] Figure 28 is a diagram showing a first configuration example of an information processing device and an information processing system including the information processing device according to Embodiment 4 of the present disclosure. [Figure 29] Figure 29 is a diagram showing a second configuration example of an information processing device and an information processing system including the information processing device according to Embodiment 4 of the present disclosure. [Figure 30] Figure 30 is a flowchart showing an example of a process for changing the number of segments in an information processing device according to Embodiment 4 of this disclosure. [Figure 31] Figure 31 is a flowchart showing an example of a process for changing the number of clusters in an information processing device according to Embodiment 4 of this disclosure. [Figure 32] Figure 32 shows a first example of a hardware configuration for realizing the functions of the configuration of this disclosure. [Figure 33] Figure 33 shows a second example of a hardware configuration for realizing the functions of the configuration of this disclosure. [Modes for carrying out the invention]

[0009] To further illustrate this disclosure, embodiments of this disclosure will be described below with reference to the accompanying drawings.

[0010] Embodiment 1. Embodiment 1 describes the form of an apparatus having the basic functions of the present disclosure and a system including said apparatus.

[0011] An example of the configuration of the information processing device according to Embodiment 1 of this disclosure will be described. Figure 1 shows an example of the configuration of an information processing device and peripheral devices related to the information processing device according to Embodiment 1 of the present disclosure. Figure 2 shows an example configuration of an information processing device and an information processing system including the information processing device according to Embodiment 1 of this disclosure. The information processing device 100 (100', 100A) according to Embodiment 1 of this disclosure changes the number of clusters in the clustering result, which is the result of dividing time-series data into segments and clustering it. The information processing device 100 (100', 100A) can also be called a time-series segmentation device from the perspective that it is a device that divides time-series data into segments and clusters them. In this disclosure, time series data can be of any type. Furthermore, time series data may be either univariate or multivariate. Multivariate time series data is, for example, multiple time series data detected simultaneously at different time points. Examples of time series data include biometric data such as voice data, electroencephalogram data, and heart rate data; sensor data such as temperature data, vibration data, power data, pressure data, weight data, and flow rate data; financial data such as stock prices, foreign exchange transactions, and power transactions; social data such as pedestrian flow data, logistics data, and traffic volume data; and environmental data such as weather data and seismic wave data. More specifically, time series data may include, for example, heart rate data acquired from wearable devices attached to the human body, vibration data acquired from industrial equipment or vehicles in operation, temperature data acquired from air conditioning or refrigeration equipment, and power data acquired from power generation equipment or power equipment. Multivariate time series data may consist of similar data (for example, multiple waveform data in electroencephalogram or electrocardiogram data) or combinations of different data (for example, temperature data and vibration data). As a concrete example of time series data, we are imagining waveform data like the time series data 1100 in Figure 10, which will be described later.

[0012] The configurations of the information processing device 100 and the information processing device 100' will be described below. The information processing device 100 shown in Figure 1 includes a change command receiving unit 110 and a cluster number changing unit 120. The information processing device of this disclosure may further include a time series division unit 101, a clustering unit 102, or a time series division unit 101 and a clustering unit 102. The information processing device 100' shown in Figure 1 includes a time series division unit 101, a clustering unit 102, a change command receiving unit 110, and a cluster number changing unit 120.

[0013] The time series splitting unit 101 splits the time series data into time series. The time series splitting unit 101 enables processing of the time series data in the divided units (segment units). The technique for splitting time series data into segments can be appropriately selected from known techniques, but in detail, for example, when the time series splitting unit 101 receives time series data, it first performs preprocessing on the time series data. There is no particular limit, but for example, it is like extracting features by time series representation learning for multivariate time series data. The time series splitting unit 101 decomposes the preprocessed time series data into partial time series (segments with N segments) based on some criterion (distance scale, etc.) based on the characteristics of the data. The criterion for splitting can be determined according to the characteristics of the data. The time series splitting unit 101 outputs the time series splitting result, which is the result of splitting the time series data into segments.

[0014] The clustering unit 102 groups the segments into clusters through clustering. The clustering unit 102 outputs clustering results, which are the result of clustering based on the characteristics of each segment using the time series partitioning results. Specifically, for example, the clustering unit 102 extracts representative points from each segment of the time series partitioning results. Representative points are, for example, the mean of each sample in the segment, the centroid, or the central sample in the segment. The clustering unit 102 clusters the obtained segment representative points with, for example, an automatically determined number of clusters K. The clustering unit 102 outputs the clustering results. The outputted clustering results are displayed, for example, on the screen of the output destination device 20. A user can then, for example, specify a new number of clusters via the input source device 10 to initiate a change.

[0015] While there are several methods for unsupervised time series segmentation, the above example illustrates a technique that decomposes time series data into segments with similar trends and then clusters each segment. Furthermore, for unsupervised time-series segmentation, techniques such as those described in Reference 1 below can also be used. References 1 Wang, Chengyu, et al. “Time2state: An unsupervised framework for inferring the latent states in time series data.” Proceedings of the ACM on Management of Data 1.1 (2023): 1-18. Reference 1 proposes a method for achieving unsupervised time series segmentation by extracting effective feature representations from univariate or multivariate time series data and clustering them using a cluster size K automatically determined by a DPGMM (Dirichlet Process Gaussian Mixture Model). However, Time2state, as described in Reference 1, automatically determines the number of clusters and does not support human feedback, so even if you want to change the number of clusters after seeing the clustering results, you cannot. Also, if you try to use a GMM-based method that allows you to specify the number of clusters, the cluster number of each sample may change significantly depending on the number of clusters given, raising concerns that consistency will not be maintained with respect to increases and decreases in the number of clusters (i.e., there is a concern that even unrelated clusters may change). In contrast, this disclosure is configured to avoid the above concerns by having the following configuration.

[0016] The change command receiving unit 110 receives a cluster number change command indicating the commanded number of clusters. The change command receiving unit 110 receives a cluster number change command, for example, in response to an operation by the user on the input source device 10. The cluster number change command includes information indicating the commanded number of clusters.

[0017] The cluster number change unit 120 performs a process to change the number of clusters in the clustering result when it receives a command. When the cluster number change unit 120 receives a command to change the number of clusters, it decides and executes the merging or splitting of clusters in the clustering result based on the clustering result and the command to change the number of clusters, and outputs the clustering result after changing to the commanded number of clusters. Specifically, the cluster number change unit 120 decides and executes the merging or splitting of clusters in the clustering result based on the clustering result output by the clustering unit 102 and the command to change the number of clusters received by the command to change unit 110, and outputs the clustering result after changing to the commanded number of clusters. In detail, for example, the cluster number change unit 120 switches between merging or splitting clusters depending on the relationship between the number of clusters K in the clustering result and the number of clusters K' (command cluster number K') indicated in the cluster number change command. If K' > K, the cluster number change unit 120 executes a process to split the clusters. If K > K', the cluster number change unit 120 executes a process to merge clusters. The clusters to be changed are automatically determined from among the clusters in the clustering result, for example, based on the characteristics of each cluster in the clustering result, so as to maintain consistency with clusters other than the clusters to be changed. The clusters to be changed may also be determined based on a user selection operation, as described in the embodiments described later (Embodiment 3, Embodiment 4). The clustering results after processing, output by the cluster count change unit 120, are output to the output destination device 20 and displayed, for example, on the screen of the output destination device 20. The cluster number modification unit 120 outputs an image in which, when a cluster is split, the new cluster resulting from the split uses colors and patterns (color, pattern, or color and pattern) that were not used in the clustering result before the change, while the other clusters that are not subject to change use the same colors and patterns (color, pattern, or color and pattern) as the clustering result before the split. The cluster number change unit 120, when merging clusters, uses a color / pattern (color, pattern, or color and pattern) that was not used in the clustering result before the change for the merged cluster, or uses the color / pattern of one of the merged clusters, while the other clusters that are not subject to change maintain the same color / pattern (color, pattern, or color and pattern) as the clustering result before the change. Furthermore, the clustering results after processing can be modified by, for example, the user specifying a new number of clusters K' (commanded number of clusters K') via the input source device 10.

[0018] Next, we will explain information processing system 1(1A). Information processing system 1(1A) is a system that includes an information processing device as shown in Figure 1. Hereinafter, the information processing system in each embodiment will be described as an example configuration comprising an information processing device 100(100') that includes a time-series division unit 101, a clustering unit 102, a change command reception unit 110, and a cluster number change unit 120. Information processing system 1(1A) can also be called a time-series segmentation system from the perspective that it is a system that includes a device for dividing time-series data into segments and clustering them. The information processing system 1A shown in Figure 2 includes an input / output device 30 and an information processing device 100 (100A). The information processing device 100 (100A) is connected to the input / output device 30 via wired or wireless communication.

[0019] The input / output device 30 shown in Figure 2 is configured to include an input source device 10 and an output destination device 20. The input / output device 30 is, for example, a terminal device that can be operated by a user for input and can output processing results to the user. The input source device 10 receives operations from the user and outputs to the information processing device 100 according to the operations. For example, it outputs data to be processed by the information processing device 100 or commands to activate the information processing device 100. The output device 20 receives data or information output from the information processing device 100. The output device 20 also outputs content to be presented to the user. The output device 20 is, for example, a display device, an analysis device, or a terminal device that includes both of these functions. The input source device 10 and the output destination device 20 may not be in the form of an integrated input / output device 30, but may be separate devices as shown in Figure 1.

[0020] The information processing device 100 receives data from the user or commands from the user via the input / output device 30. The information processing device 100 also outputs the processing results to the user via the input / output device 30.

[0021] The information processing device (100) 100A shown in Figure 2 is configured to include a time-series division unit 101, a clustering unit 102, a change command reception unit 110, and a cluster number change unit 120. The time-series division unit 101 and the clustering unit 102 are configured in the same manner as the information processing device 100' described earlier. The change command receiving unit 110 is configured in detail to include a cluster number change command receiving unit 111, as shown in Figure 2, for example. The cluster number change command receiving unit 111 receives a cluster number change command indicating the number of clusters commanded by the user via the input source device 10 of the input / output device 30.

[0022] The cluster number modification unit 120, in more detail, for example, merges or splits the clusters to be modified, while maintaining the clusters other than those to be modified. Figure 3 shows an example of the configuration of the cluster number changing unit in the information processing device according to Embodiment 1 of this disclosure. The cluster number changing unit 120 shown in Figure 3 is composed of a cluster merging unit 121 and a cluster splitting unit 122. The cluster merging unit 121 of the cluster number changing unit 120 merges multiple clusters that have been determined to be changed by the cluster number changing unit 120. The cluster merging unit 121 merges the clusters that are to be changed from the clustering results and maintains the clusters that are not to be changed. The cluster division unit 122 of the cluster number change unit 120 divides the clusters that have been determined to be changed by the cluster number change unit 120. The cluster division unit 122 divides the clusters that are to be changed from the clustering results and maintains the clusters that are not to be changed.

[0023] The information processing device 100 (100', 100A) may also include, in addition to the above configuration, a control unit (not shown), a storage unit (not shown), and a communication unit (not shown). A control unit (not shown) controls the entire information processing device 100 (100', 100A) and its individual components. For example, the control unit starts up the information processing device 100 (100', 100A) according to an external command. The control unit also controls the state of the information processing device 100 (operating state = state such as startup, shutdown, sleep, etc.). A storage unit (not shown) stores the data used by the information processing device 100 (100', 100A). For example, the storage unit (not shown) stores the output (output data) from each component of the information processing device 100 (100', 100A) and outputs the requested data to the requesting component. The communication unit (not shown) communicates with external devices. For example, it communicates between the information processing device 100 (100A) and peripheral devices (for example, external devices other than the information processing device 100 (100', 100A)). For example, if the information processing device 100 (100', 100A) and the external input / output device 30 are not connected by wire, the communication unit (not shown) has the function of communicating between the information processing device 100 (100', 100A) and the input / output device 30. The communication unit (not shown) may also be configured to have the function of communicating with an external server device. The control unit (not shown), storage unit (not shown), and communication unit (not shown) are the same in the embodiments described later.

[0024] Next, an example of processing by the information processing device according to Embodiment 1 of this disclosure will be described. The processing performed by an information processing device is the method of information processing performed by the information processing device. The processing performed by the information processing device is executed, for example, by a program that causes a computer to operate as such an information processing device. The processing of information processing devices 100, 100', and 100A differs in that information processing device 100A explicitly states that the source of information used in processing is the input source device, and information processing device A explicitly states that the destination of the processed data is the output destination device. However, since the internal processing is similar for each, an example of the processing of information processing device 100A will be explained here as a representative example.

[0025] Figure 4 is a flowchart showing an example of processing in an information processing device according to Embodiment 1 of this disclosure. The information processing device (time-series segmentation device) 100 (100A) according to this embodiment, for example, when it receives time-series data, starts the processing shown in Figure 4 ("Start").

[0026] The information processing device 100A then performs time-series partitioning (step ST1010 “time-series partitioning”). In the time-series partitioning process, the time-series partitioning unit 101 of the information processing device 100A partitions the time-series data into time-series segments. The time-series partitioning unit 101 enables processing of the partitioned units (segment units) of the time-series data. The time-series partitioning unit 101 outputs the time-series partitioning results, including information for each segment, to the clustering unit 102.

[0027] Here, we will explain a detailed example of the time series splitting process (step ST1010 "time series splitting"). Figure 5 is a flowchart showing a detailed example of the time-series division process in the processing of the information processing device according to Embodiment 1 of this disclosure. The information processing device 100A according to this embodiment executes the process shown in Figure 5 during the time-series division process of step ST1010 shown in Figure 4. When the time series splitting unit 101 of the information processing device 100A according to this embodiment starts the process shown in Figure 5 ("Start"), it then performs preprocessing (step ST1110 "Preprocessing of time series data"). For example, it performs time series representation learning on multivariate time series data to derive features. The time series splitting unit 101 then performs a segmentation process (step ST1120 "Segmentation"). The time series splitting unit 101 decomposes the preprocessed time series data into partial time series (segments with N segments) based on criteria (distance scale, etc.) that are based on the characteristics of the data. The time-series splitting unit 101 then performs time-series splitting result output processing (step ST1130 “Time-series splitting result output”). The time-series splitting unit 101 outputs the time-series splitting results, including information for each segment, to the clustering unit 102. Alternatively, the time-series splitting unit 101 may output the time-series splitting results, including segment information, to the output destination device 20 (see, for example, the embodiments described later). When the time series splitting unit 101 executes the time series splitting result output processing, it terminates the series of processes shown in Figure 5 ("terminate"). The time series splitting unit 101 continues to perform the series of processes shown in Figure 5, for example, until the input of time series data is completed, or for example, until it receives a termination command from an external source.

[0028] Returning to the explanation of Figure 4. Following the time-series partitioning process in step ST1010, the information processing device 100A performs a clustering process (step ST1020 "clustering"). In the clustering process, the clustering unit 102 of the information processing device 100A outputs a clustering result, which is the result of clustering based on the characteristics of each segment using the time-series partitioning result.

[0029] Here, we will describe a detailed example of the clustering process (step ST1020 "Clustering"). Figure 6 is a flowchart showing a detailed example of clustering processing in the information processing apparatus according to Embodiment 1 of this disclosure. The information processing device 100A according to this embodiment executes the process shown in Figure 6 during the clustering process of step ST1020 shown in Figure 4. When the clustering unit 102 of the information processing device 100A according to this embodiment receives the time-series division result output by the time-series division unit 101, it starts the process shown in Figure 6 ("Start"). The clustering unit 102 then performs a segment representative point extraction process (step ST1210 "Extraction of representative points for each segment"). In the segment representative point extraction process, the clustering unit 102 calculates and extracts representative points such as the mean, centroid, and central sample within the segment for each sample in the segment. The clustering unit 102 then executes the clustering execution process (step ST1220 "Execute clustering with K clusters"). The clustering unit 102 clusters the obtained segment representative points, for example, with an automatically determined number of clusters K. The clustering unit 102 then performs clustering result output processing (step ST1230 "Clustering Result Output"). The clustering unit 102 outputs the clustering results to the output destination device 20. When the clustering unit 102 executes the clustering result output processing, it terminates the series of processes shown in Figure 6 ("terminate"). The clustering unit 102 continues to perform the series of processes shown in Figure 6, for example, until the input of time-series data is completed, or for example, until it receives a termination command from an external source.

[0030] Returning to the explanation of Figure 4. After the information processing device 100A executes the clustering process in step ST1020, it then executes the change command reception determination process (step ST1030 "Change command received?"). In the change command reception determination process, the change command reception unit 110 of the information processing device 100A makes a determination based on the result of confirming whether a cluster number change command has been received. Specifically, for example, the cluster number change command reception unit 111 of the change command reception unit 110 makes the determination and outputs the cluster number change command and the determination result to the cluster number change unit 120. If the cluster number change command reception unit 111 does not receive a cluster number change command within a preset time period, it determines that a cluster number change command has not been received. As previously explained, the cluster number change command includes information indicating the number of clusters commanded by the user via the input source device 10.

[0031] If the information processing device 100A determines in the change command reception determination process of step ST1030 that it has received a cluster number change command (step ST1030 "Change command received?" "YES"), it then executes the cluster number change process (step ST1040 "Change cluster number"). In the cluster number change process, the cluster number change unit 120 of the information processing device 100A decides and executes the merging or splitting of clusters in the clustering result based on the clustering result and the cluster number change command. When the information processing device 100A executes the cluster number change process in step ST1040, it then executes the clustering result output process (step ST1050 "Output clustering results"). In the clustering result output process, the cluster number change unit 120 of the information processing device 100A outputs the clustering results after the cluster number has been changed. After the information processing device 100A executes the clustering result output processing in step ST1050, it then proceeds to step ST1030 to execute the change command acceptance determination processing.

[0032] If the information processing device 100A determines in the change command reception determination process of step ST1030 that it has not received a command to change the number of clusters (step ST1030 "Change command received?" "NO"), it then executes a termination determination process (step ST1060 "Termination?"). In the termination determination process, the control unit (not shown) of the information processing device 100A determines to terminate the process, for example, if it determines that the input of time-series data has been completed. Alternatively, for example, it determines to terminate the process if it receives a termination command from an external source. If the information processing device 100A determines in the termination determination process of step ST1060 that it is not terminating (step ST1060 "Termination?" "NO"), it returns to the process of step ST1010. If the information processing device 100A determines to terminate in the termination determination process of step ST1060 (step ST1060 "Termination?" "YES"), it terminates the process shown in Figure 4 ("Termination").

[0033] Here, we will explain a detailed example of the process of changing the number of clusters in an information processing device. Figure 7 is a flowchart showing a detailed example of the process for changing the number of clusters in the information processing device according to Embodiment 1 of this disclosure. The information processing device 100A according to this embodiment executes the process shown in Figure 7, for example, in the change command receiving unit 110 and the cluster number change unit 120. The information processing device 100A starts the process shown in Figure 7 ("Start"), following, for example, step ST1020 shown in Figure 4. The information processing device 100A then executes a cluster number change command reception determination process (step ST1310 "Cluster number change command received?"). In this process, the cluster number change command reception unit 111 of the change command reception unit 110 in the information processing device 100A determines whether it has received a cluster number change command indicating the commanded number of clusters, and outputs the cluster number change command and the determination result to the cluster number change unit 120. The cluster number change unit 120 of the information processing device 100A then executes a cluster number change process (step ST1320 "Cluster number change"). In the cluster number change process, the cluster number change unit 120 decides and executes the merging or splitting of clusters in the clustering result based on the clustering result and the cluster number change command. The cluster number change unit 120 of the information processing device 100A then performs clustering result output processing (step ST1330 "Output clustering results"). The cluster number change unit 120 outputs the processed clustering results to the output destination device 20. After executing the process in step ST1330, or if the information processing device 100A determines in step ST1310 that it did not receive a command to change the number of clusters (step ST1310 "Command to change the number of clusters received?" "NO"), it then executes a termination determination process (step ST1340 "Termination?"). In the termination determination process, the control unit (not shown) of the information processing device 100A determines to terminate the process, for example, if it determines that the input of time-series data has been completed. Alternatively, for example, it determines to terminate the process if it receives a termination command from an external source. If the information processing device 100A determines in the termination determination process of step ST1340 that it is not terminating (step ST1340 "Termination?" "NO"), it returns to the process of step ST1310. If the information processing device 100A determines to terminate in the termination determination process of step ST1340 (step ST1340 "Termination?" "YES"), it terminates the process shown in Figure 7 ("Termination").

[0034] A more detailed example of the process shown in Figure 7 will be explained. Figure 8 is a flowchart showing a more detailed example of the process shown in Figure 7. The information processing device 100A according to this embodiment executes the process shown in Figure 8, for example, in the change command reception unit 110 and the cluster number change unit 120. When the information processing device 100A starts the process shown in Figure 8 ("Start"), the cluster number change command receiving unit 111 of the change command receiving unit 110 in the information processing device 100A then executes a cluster number change command reception determination process (step ST1410 "Cluster number change command received?"). This process is the same as the cluster number change command reception determination process in step ST1310. If the cluster number change unit 120 of the information processing device 100A determines in the cluster number change command reception determination process of step ST1410 that it has received a cluster number change command (step ST1410 "Cluster number change command received?" "YES"), it then executes a clustering result acquisition process (step ST1420 "Acquire clustering results"). In this process, the cluster number change unit 120 acquires the clustering results output by the clustering unit 102. The cluster number change unit 120 of the information processing device 100A then performs a change type determination process (step ST1430 "Determine whether to merge or split clusters based on the change command and clustering results"). In this process, the cluster number change unit 120 decides whether to merge or split the clusters, which are the clustering results, based on the clustering results output by the clustering unit 102 and the cluster number change command received by the change command reception unit 110. Specifically, for example, the cluster number change unit 120 decides to switch between merging or splitting clusters depending on the relationship between the number of clusters K in the clustering results and the command cluster number K' indicated in the cluster number change command. If the cluster number change unit 120 of the information processing device 100A decides to merge clusters in the process of step ST1430 (step ST1430 "Deciding whether to merge or split clusters based on the change command and clustering results" "Merge"), it then executes the cluster merging process (step ST1440 "Cluster Merging"). The cluster merging unit 121 of the cluster number change unit 120 merges multiple clusters that are subject to change and have similar characteristics in the clustering results into one cluster to form the changed cluster. If the cluster number change unit 120 of the information processing device 100A decides to split a cluster in the process of step ST1430 (step ST1430 "Deciding whether to merge or split clusters based on change command and clustering results" "Split"), it then executes a cluster splitting process (step ST1450 "Cluster splitting"). In the cluster splitting process, the cluster splitting unit 122 of the cluster number change unit 120 splits the cluster to be changed in a way that has the least impact on clusters other than the cluster to be changed among the clusters in the clustering results. The cluster number changing unit 120 of the information processing device 100A then performs clustering result output processing (step ST1460 "Output clustering results"). This processing is the same as the clustering result output processing in step ST1330. After executing the process in step ST1460, or if the cluster number change unit 120 of the information processing device 100A determines in step ST1410 that it did not receive a cluster number change command (step ST1410 "Cluster number change command received?" "NO"), it then executes a termination determination process (step ST1470 "Termination?"). This process is the same as the termination determination process in step ST1340. If the information processing device 100A determines in the termination determination process of step ST1470 that it is not terminating (step ST1470 "Termination?" "NO"), it returns to the process of step ST1410. If the information processing device 100A determines to terminate in the termination determination process of step ST1470 (step ST1470 "Termination?" "YES"), it terminates the process shown in Figure 8 ("Termination").

[0035] Here, we will describe a more detailed example of the process in steps ST1430, ST1440, and ST1450 shown in Figure 8. Figure 9 is a flowchart showing a more detailed example of the process for changing the number of clusters in the information processing device according to Embodiment 1 of this disclosure. The information processing device 100A according to this embodiment executes the process shown in Figure 9, for example, in the cluster number changing unit 120. The information processing device 100A according to this embodiment 1 executes the process shown in Figure 9, following the process of step ST1420 (clustering result acquisition process) shown in Figure 8. The cluster number change unit 120 of the information processing device 100A then executes a command cluster number acquisition process (step ST1510 "Acquire cluster number K'"). In this process, the cluster number change unit 120 acquires the cluster number K' (command cluster number K') indicated in the cluster number change command output by the cluster number change command reception unit 111. The cluster number change unit 120 then executes a cluster number change determination process (step ST1520 "K'!=K?"). In this process, the cluster number change unit 120 determines whether the cluster number K of the clustering result matches the command cluster number K' indicated in the cluster number change command. If the cluster number K of the clustering result does not match the command cluster number K' indicated in the cluster number change command, the cluster number change unit 120 determines to change the cluster number. The cluster number change unit 120 determines that it will not change the cluster number if the cluster number K of the clustering result does not match the command cluster number K' indicated in the cluster number change command (step ST1520 "K'!=K?" "YES"), and proceeds to the cluster result output processing in step ST1460, as shown in Figure 8. The cluster number change unit 120 determines to change the cluster number if the cluster number K of the clustering result does not match the command cluster number K' indicated in the cluster number change command (step ST1520 "K'!=K?" "NO"), and then executes a change type determination process (step ST1530 "K'>K?"). In this process, the cluster number change unit 120 determines to execute a cluster splitting process if the command cluster number K' indicated in the cluster number change command is greater than the cluster number K of the clustering result. The cluster number change unit 120 determines to execute a cluster merging process if the command cluster number K' indicated in the cluster number change command is less than the cluster number K of the clustering result. If the cluster number change unit 120 determines that the command cluster number K' indicated in the cluster number change command is greater than the cluster number K in the clustering result (step ST1530 "K'>K?" "YES"), it then executes a cluster partitioning process (step ST1540 "Partition clusters K'-K times"). In this process, the cluster partitioning unit 122 of the cluster number change unit 120 executes a process K'-K times to partition the cluster to be changed from among the clusters in the clustering result in such a way that it has the least impact on clusters other than the cluster to be changed. If the number of clusters K' indicated in the cluster number change command is smaller than the number of clusters K in the clustering result (step ST1530 "K'>K?" "NO"), the cluster number change unit 120 then performs a cluster merging process (step ST1550 "Merge clusters KK' times"). In this process, the cluster merging unit 121 of the cluster number change unit 120 performs a process KK' times to merge multiple clusters that are the target of modification, which have similar characteristics and have the least impact on clusters other than the target cluster, into one of the clusters to be modified. When the information processing device 100A according to this embodiment 1 executes the cluster partitioning process in step ST1540 or the cluster merging process in step ST1550, it then terminates the series of processes shown in Figure 9 ("termination") and proceeds to the clustering result output process shown in Figure 8 (step ST1460 "output clustering results").

[0036] Here, we will explain time-series partitioning and clustering. Figure 10 shows an illustration of unsupervised time series segmentation and clustering. Figure 10 shows an example where time-series data 1100 is first decomposed into six segments 1200, and then clustering is performed to obtain five clusters 1300 (the two white segments belong to the same cluster). The time series splitting unit 101 extracts features of the time series data 1100 as shown in Figure 10, and decomposes it into segments 1200 as shown in Figure 10, for each partial time series that show similar trends based on criteria such as a distance scale based on the features. The clustering unit 102 extracts a representative point for each of the 1200 segments as shown in Figure 10, and performs clustering by classifying the segments 1200 into clusters 1300 with a cluster count of K using the values ​​of the representative points.

[0037] Figure 11 shows an image illustrating how to change the number of clusters in the information processing device of this disclosure. The left side of Figure 11 shows an example of clustering results before changing the number of clusters, and the right side of Figure 11 shows an example of clustering results after changing the number of clusters. For the sake of clarity, Figure 11 shows an example of changing the number of clusters from 3 to 2. When the change command receiving unit 110 of the information processing device 100 receives a cluster number change command with a command cluster number of 2 for a clustering result of 3 clusters (cluster #1, cluster #2, cluster #3) as shown on the left side of Figure 11, the cluster number change unit 120 of the information processing device 100 decides to execute a cluster merging process once based on the clustering result and the cluster number change command, since cluster number 3 > command cluster number 2. Based on the characteristics of cluster #1, cluster #2, and cluster #3, cluster #2 and cluster #3 are selected as change candidates, and cluster #3 is merged into cluster #2 as shown on the right side of Figure 11.

[0038] Figure 12 shows an image of what happens when you try to change the number of clusters in a configuration that does not employ the information processing device of this disclosure. The left side of Figure 12 shows an example of clustering results before changing the number of clusters, and the right side of Figure 12 shows an example of clustering results after hypothetically changing the number of clusters in a configuration that does not employ the information processing device of this disclosure. If we consider the clustering result with 3 clusters (cluster #1, cluster #2, cluster #3) on the left side of Figure 12 and attempt to change the number of clusters from 3 to 2, and decide to target cluster #3, it is possible that a portion of cluster #3 will be clustered into cluster #1, and another portion of cluster #3 will be clustered into cluster #2 (in the example shown in Figure 12, the portion corresponding to cluster #3 is split and clustered alternately into cluster #1 and cluster #2). When clustering occurs in this way, cluster #3 in the original clustering result becomes fragmented, lacking consistency and making analysis difficult.

[0039] Next, an example of the processing results and user interface output via an output destination device by the information processing device of this disclosure will be described. Figure 13 shows an example of a user interface (image) for an output destination device output from the information processing device of this disclosure (an example before changing the number of clusters). Figure 14 shows an example of a user interface (image) for an output destination device output from the information processing device of this disclosure (an example after changing the number of clusters). The output image (screen) 1500 shown in Figure 13 includes time-series data 1510, clustering results 1520, cluster count command operation unit 1530, command cluster count display unit 1531, and clustering target data selection unit 1540. The output image (screen) 1600 shown in Figure 14 includes time-series data 1610, clustering results 1620, cluster count command operation unit 1630, command cluster count display unit 1631, and clustering target data selection unit 1640.

[0040] The time series data 1510 and clustering results 1520 shown in Figure 13 are displayed in a way that corresponds to the time series. Similarly, the time series data 1610 and clustering results 1620 shown in Figure 14 are displayed in a way that corresponds to the time series. Although not shown in the figures, if the time series data is multivariate time series data, the system may be configured to display each time series data in a way that corresponds to the time series. This clearly shows the correspondence between the waveform of the time series data and the segments and clusters, making it easy for the user to understand.

[0041] The cluster number command operation unit 1530 is configured to allow the command number of clusters to be changed by sliding a slider. The command cluster number display unit 1531 is configured to display the command cluster number in response to operations performed on the cluster number command operation unit 1530. The cluster number command operation unit 1630 and the command cluster number display unit 1631 are configured similarly. This configuration makes it easier for the user to perform operations.

[0042] The clustering target data selection unit 1540 and the clustering target data selection unit 1640 are configured to allow users to change the time series data to be displayed. For example, if the time series data is multivariate time series data, they are configured to allow users to select and display each variable separately.

[0043] Furthermore, if cluster 3 is merged into cluster 2 in the clustering result shown in Figure 13, changing the result to the clustering result shown in Figure 14, the color of cluster 2 (cluster 2 after merging) in the clustering result shown in Figure 14 will be maintained at the same color as cluster 2 (cluster 2 before merging) in Figure 13. This visually demonstrates the consistency of the clustering results to the user.

[0044] In clustering, there is a need in practical applications to visualize the clustering results and interactively change the number of clusters, thereby aligning the user's domain knowledge with the actual clustering. According to the information processing device of Embodiment 1 of this disclosure, it is possible to satisfy the above-mentioned needs.

[0045] When comparing this disclosure with Patent Document 1 described at the beginning, Patent Document 1 (JP 2024-142699 A (Yanmar)) uses an indicator called "granularity," which is not directly linked to the visualization results, making it difficult to reconcile the user's domain knowledge with the actual clustering. Furthermore, the technology in Patent Document 1 does not guarantee consistency before and after the change. In addition, because clustering is reapplied to inappropriate cluster results, from an application perspective, high computational costs result in waiting times, which impairs the user experience. In contrast, the information processing device with the configuration described above, as disclosed in this disclosure, allows for the direct visualization and direct modification of the number of clusters in time-series segmentation tasks. Furthermore, it enables consistent clustering in response to increases or decreases in the number of clusters. Additionally, it reduces the waiting time when increasing or decreasing the number of clusters. As a result, the user experience when changing the number of clusters can be improved.

[0046] This embodiment shows an example of a configuration that can be expressed as follows. [1] An information processing device that modifies clustering results, which are obtained by dividing time-series data into segments and clustering them, A change command receiving unit that receives a command to change the number of clusters, which indicates the commanded number of clusters, A cluster number change unit, upon receiving the aforementioned cluster number change command, determines and executes the merging or splitting of clusters in the clustering result based on the clustering result and the cluster number change command, and outputs the clustering result after the cluster number has been changed. An information processing device characterized by having the following features. This disclosure has the effect of providing an information processing device that enables the maintenance of consistency between the clustering before and after a change when modifying the clustering results. Furthermore, by applying the above-mentioned information processing device to a time-series segmentation device, it is possible to provide a time-series segmentation device that, similarly to the above, enables the maintenance of clustering consistency when modifying clustering results.

[0047] This embodiment shows an example of a configuration that can be expressed as follows.

[10] An information processing method by an information processing device that modifies clustering results, which are the result of clustering time-series data, The change command receiving unit of the information processing device receives a cluster number change command indicating the commanded number of clusters, When the cluster number change command is received, the cluster number change unit of the information processing device decides and executes the merging or splitting of clusters in the clustering result based on the clustering result and the cluster number change command, and outputs the clustering result after the cluster number has been changed. An information processing method characterized by the following: This disclosure has the effect of providing an information processing method that enables maintaining the consistency of clustering before and after a change when modifying the clustering results.

[0048] This embodiment shows an example of a configuration that can be expressed as follows.

[11] Computers, An information processing device that modifies clustering results, which are the result of clustering time-series data, A change command receiving unit that receives a command to change the number of clusters, which indicates the commanded number of clusters, A cluster number change unit, upon receiving the aforementioned cluster number change command, determines and executes the merging or splitting of clusters in the clustering result based on the clustering result and the cluster number change command, and outputs the clustering result after the cluster number has been changed. A program characterized by operating as an information processing device equipped with [a specific feature]. This means that the disclosure can provide a program that enables maintaining the consistency of the clustering before and after a change when modifying the clustering results.

[0049] This embodiment further illustrates an example of a configuration that can be expressed as follows. [2] The aforementioned cluster number changing unit is: A cluster partitioning unit divides the clusters to be modified from the aforementioned clustering results and maintains the clusters other than those to be modified. A cluster merging unit merges the clusters to be changed from the aforementioned clustering results, while maintaining the clusters that are not to be changed. Equipped with, The information processing apparatus according to [1], characterized in that This disclosure provides an information processing device that enables the maintenance of clustering consistency, as it ensures the consistency of clusters other than those being modified when modifying the clustering results. Furthermore, by applying the above-mentioned information processing device to a time-series segmentation device, it is possible to provide a time-series segmentation device that enables the maintenance of clustering consistency when modifying clustering results, similar to the above. Furthermore, by applying the above configuration to a system including an information processing device, the above information processing method, or the above program, etc., this disclosure will produce the same effects as described above.

[0050] This embodiment further illustrates an example of a configuration that can be expressed as follows. [9] A time-series segmentation unit that divides the aforementioned time-series data into segments, A clustering unit that outputs the clustering result, which is the result of clustering based on the respective characteristics of each segment, Equipped with, The cluster number change unit determines and executes the merging or splitting of the clusters, which are the clustering results, based on the clustering results output by the clustering unit and the cluster number change command, and outputs the clustering results after the cluster number has been changed. An information processing device according to any one of [1][2][3][4][5][6][7][8], characterized by the above. This disclosure provides an information processing device that takes time-series data as input and outputs clustering results, and that enables the maintenance of clustering consistency when modifying the clustering results. Furthermore, by applying the above-mentioned information processing device to a time-series segmentation device, it is possible to provide a time-series segmentation device that enables the maintenance of clustering consistency when modifying clustering results, similar to the above. Furthermore, by applying the above configuration to a system including an information processing device, the above information processing method, or the above program, etc., this disclosure will produce the same effects as described above.

[0051] Embodiment 2. In the above-described embodiment 1, an example of a configuration in which clusters are divided or merged while maintaining the consistency of the clusters in the clustering results was explained. Embodiment 2 further describes an example of a configuration in which segments are divided or merged while maintaining consistency. In Embodiment 2, components of Embodiment 2 that are the same as those of Embodiment 1 already described are denoted by the same reference numerals or similar numerals, and redundant explanations are omitted as appropriate.

[0052] Next, an example of the configuration of the information processing device according to Embodiment 2 of this disclosure will be described. Figure 15 shows an example configuration of an information processing device and an information processing system including the information processing device according to Embodiment 2 of this disclosure. The information processing device (time-series segmentation device) 100 (100B) and the information processing system (time-series segmentation system) 1 (1B) including the information processing device according to this second embodiment mainly differ from the configuration of the embodiment described above in that it has a function that allows for further changes to the number of segments.

[0053] The information processing device 100 (100B) shown in Figure 15 is composed of a time-series division unit 101, a clustering unit 102, a change command reception unit 110, a cluster number change unit 120, and a segment number change unit 130.

[0054] The time series splitting unit 101, similar to the time series splitting unit already described, splits the time series data into segments (number of segments N) and outputs the time series splitting result. The output time-series segmentation results are displayed, for example, on the screen of the output destination device 20. A user can then, for example, specify a new number of segments N' (command number of segments N') via the input source device 10 to issue a change command.

[0055] The change command receiving unit 110 receives a cluster number change command indicating the commanded number of clusters. The change command receiving unit 110 also receives a segment number change command indicating the commanded number of segments. The change command receiving unit 110 shown in Figure 15 is composed of a cluster number change command receiving unit 111 and a segment number change command receiving unit 112. The segment number change command receiving unit 112 receives a segment number change command indicating the commanded number of segments. The segment number change command receiving unit 112 receives a segment number change command, for example, in response to an operation by the user on the input source device 10. The cluster number change command receiving unit 111 receives a cluster number change command indicating the commanded number of clusters. The cluster number change command receiving unit 111 receives a cluster number change command that is commanded in response to, for example, an operation by the user on the input source device 10.

[0056] The segment number changing unit 130, upon receiving a command, performs a process to change the number of segments in the time-series division result. When the segment number change unit 130 receives a segment number change command, it decides and executes the merging or splitting of the segments, which are the result of dividing the time series data into segments, based on the time series division result and the segment number change command, and outputs the time series division result after changing the number of segments. Specifically, the segment number change unit 130 determines and executes the merging or splitting of segments, which are the time series division results, based on the time series division results output by the time series division unit 101 and the segment number change command received by the segment number change command receiving unit 112 of the change command receiving unit 110, and outputs the time series division results after changing to the commanded number of segments. In detail, for example, the segment number change unit 130 switches between merging or splitting segments depending on the relationship between the number of segments N in the time-series division result and the number of segments N' (command segment number N') indicated in the segment number change command. If N' > N, the segment number change unit 130 executes a segment splitting process. If N > N', the segment number change unit 130 executes a segment merging process. The segments to be changed are automatically determined, for example, based on the characteristics of each segment in the time-series division result, so as to maintain consistency with segments other than the segments to be changed from among the segments in the time-series division result. The segments to be changed may also be determined based on a user selection operation, as described in the embodiments described later (Embodiment 3, Embodiment 4). The processed time-series division results output by the segment number change unit 130 are output to the output destination device 20 and displayed, for example, on the screen of the output destination device 20. When a segment is split, the segment number changing unit 130 outputs an image in which the new segments created by the split use colors and patterns (color, pattern, or color and pattern) that were not used in the time-series splitting result before the change, while the other segments that are not subject to change use the same colors and patterns (color, pattern, or color and pattern) as the time-series splitting result before the change. When segments are merged, the segment number changing unit 130 uses a color / pattern (color, pattern, or color and pattern) that was not used in the time-series division result before the change, or uses the color / pattern of one of the merged segments, while the other segments that are not subject to change maintain the same color / pattern (color, pattern, or color and pattern) as the time-series division result before the change. In contrast, for example, a user can also specify a new number of segments N' (command number of segments N') via the input source device 10 to issue a change command.

[0057] The cluster number changing unit 120 has the same functions as the cluster number changing unit 120 according to the embodiment described above. In other words, the cluster number change unit 120 performs a process to change the number of clusters in the clustering result when it receives a command. Specifically, when the cluster number change unit 120 receives a command to change the number of clusters, it decides and executes the merging or splitting of clusters in the clustering result based on the clustering result and the command to change the number of clusters, and outputs the clustering result after changing to the commanded number of clusters. The clustering results after processing are displayed on the screen of the output device 20, for example, in the same way as the clustering results before the change. Alternatively, a user can specify a new number of clusters K' (commanded number of clusters K') via the input device 10 to issue a change command. Note that "after processing" may include not only the results after processing when a change is made, but also the results after processing when no change is made.

[0058] Here, we will describe a detailed configuration example of the segment number changing unit 130 and the cluster number changing unit 120. Figure 16A shows an example of the configuration of the segment number changing unit in the information processing device according to Embodiment 2 of the present disclosure, and Figure 16B shows an example of the configuration of the cluster number changing unit in the information processing device according to Embodiment 2 of the present disclosure.

[0059] The segment number changing unit 130 shown in Figure 16A is composed of a segment merging unit 131 and a segment splitting unit 132. The segment merging unit 131 merges the segments to be changed from the clustering results, while maintaining the segments that are not to be changed. If the segment number change unit 130 switches to the merging process, it executes the process. The segment division unit 132 divides the segments to be changed from the time-series division results and maintains the segments that are not to be changed. The segment division unit 132 executes the process when the segment number change unit 130 switches to division processing.

[0060] The cluster number changing unit 120 shown in Figure 16B is composed of a cluster merging unit 121 and a cluster splitting unit 122. The cluster merging unit 121 merges the clusters to be modified from the clustering results, while maintaining the clusters that are not to be modified. The cluster merging unit 121 executes the process when the cluster number modification unit 120 switches to the merging process. The cluster partitioning unit 122 partitions the clusters to be modified from the clustering results, while maintaining the clusters that are not to be modified. The cluster partitioning unit 122 executes the process when the cluster number modification unit 120 switches to partitioning processing.

[0061] Next, an example of processing by the information processing device according to Embodiment 2 of this disclosure will be described. Figure 17 is a flowchart showing an example of processing in an information processing device according to Embodiment 2 of this disclosure. The information processing device (time-series segmentation device) 100 (100B) according to this embodiment, for example, when it receives time-series data, starts the processing shown in Figure 17 ("Start"). The information processing device 100B then performs time-series partitioning (step ST2010 “Time-series partitioning”). In the time-series partitioning process, the time-series partitioning unit 101 of the information processing device 100B partitions the time-series data into segments. The time-series partitioning unit 101 divides the time-series data into segments. The time-series partitioning unit 101 outputs the time-series partitioning results, including information for each segment, to the clustering unit 102. The details of the time-series partitioning process are the same as those shown in Figure 5, which has already been explained, and a detailed explanation is omitted here.

[0062] After executing the time-series division process in step ST2010, the information processing device 100B then executes a segment number change command reception determination process (step ST2020 "Segment number change command received?"). In this process, the change command reception unit 110 of the information processing device 100B makes a determination based on the result of confirming whether a segment number change command has been received. Specifically, for example, after the time-series division result is output, the segment number change command reception unit 112 of the change command reception unit 110 checks whether a segment number change command has been received for a predetermined period of time and makes a determination, outputting the segment number change command and the determination result to the segment number change unit 130. If the segment number change command reception unit 112 does not receive a segment number change command during the predetermined period of time, it determines that a segment number change command was not received. As previously explained, the segment number change command includes information indicating the number of segments commanded by the user via the input source device 10.

[0063] If the information processing device 100B determines in the segment number change command reception determination process of step ST2020 that it has received a segment number change command (step ST2020 "Cluster number change command received?" "YES"), it then executes the segment number change process (step ST2030 "Segment number change"). In the segment number change process, the segment number change unit 130 of the information processing device 100B decides and executes the merging or splitting of segments in the time series division result based on the time series division result and the segment number change command.

[0064] When the information processing device 100B executes the segment number change process in step ST2030, it then executes the time series division result output process (step ST2040 "Output time series division result"). In the time series division result output process, if the segment number change unit 130 of the information processing device 100B changes the segment number, it outputs the time series division result after the segment number change to the output destination device 20. The segment number change unit 130 also outputs the time series division result to the clustering unit 102.

[0065] The information processing device 100B then performs clustering (step ST2050 "Clustering"). In the clustering process, the clustering unit 102 of the information processing device 100B outputs clustering results, which are the result of clustering based on the characteristics of each segment using the time-series partitioning results. The details of the clustering process are the same as those of the clustering process in Figure 6, which has already been described, and a detailed explanation is omitted here.

[0066] After the information processing device 100B executes the clustering process in step ST2050, it then executes the cluster number change command reception determination process (step ST2060 "Cluster number change command received?"). In this process, the change command reception unit 110 of the information processing device 100B makes a determination based on the result of confirming whether a cluster number change command has been received. Specifically, for example, after the clustering result is output to the output destination device 20, the cluster number change command reception unit 111 of the change command reception unit 110 checks whether a cluster number change command has been received for a predetermined period of time and makes a determination, and outputs the cluster number change command and the determination result to the cluster number change unit 120. If the cluster number change command reception unit 111 does not receive a cluster number change command during the predetermined period of time, it determines that a cluster number change command was not received. As previously explained, the cluster number change command includes information indicating the number of clusters commanded by the user via the input source device 10.

[0067] If the information processing device 100B determines in the cluster number change command reception determination process in step ST2060 that it has received a cluster number change command (step ST2060 "Cluster number change command received?" "YES"), it then executes the cluster number change process (step ST2070 "Cluster number change"). In the cluster number change process, the cluster number change unit 120 of the information processing device 100B decides and executes the merging or splitting of clusters in the clustering result based on the clustering result and the cluster number change command.

[0068] The information processing device 100B then performs clustering result output processing (step ST2080 "Output clustering results"). In the clustering result output processing, if the cluster number change unit 120 of the information processing device 100B changes the cluster number, it outputs the clustering results after the cluster number change to the output destination device 20. The output destination device 20 displays, for example, an image including the clustering results.

[0069] After the information processing device 100B executes the clustering result output processing in step ST2080, it then proceeds to step ST2060 to execute the change command acceptance determination processing. If the information processing device 100B determines in the change command reception determination process of step ST2060 that it has not received a command to change the number of clusters (step ST2060 "Change command received?" "NO"), it then executes a termination determination process (step ST2090 "Termination?"). In the termination determination process, the control unit (not shown) of the information processing device 100B determines to terminate the process, for example, if it determines that the input of time-series data has been completed. Alternatively, for example, it determines to terminate the process if it receives a termination command from an external source. If the information processing device 100B determines in the termination determination process of step ST2090 that it is not terminating (step ST2090 "Termination?" "NO"), it returns to the process of step ST2010. If the information processing device 100B determines to terminate in the termination determination process of step ST2090 (step ST2090 "Termination?" "YES"), it terminates the process shown in Figure 17 ("Termination").

[0070] Next, we will explain a detailed example of the process of changing the number of segments in an information processing device. Figure 18 is a flowchart showing a more detailed example of the process of changing the number of segments, which is one of the processes shown in Figure 17. The information processing device 100B according to this embodiment executes the process shown in Figure 18, for example, in the change command receiving unit 110 and the segment number changing unit 130. When the information processing device 100B starts the process shown in Figure 18, it then executes a segment number change command reception determination process (step ST2310 "Segment number change command received?"). In the segment number change command reception determination process, the segment number change command reception unit 112 in the change command reception unit 110 of the information processing device 100B checks whether a segment number change command has been received and makes a determination, similar to the process in step ST2020 already described, and outputs the segment number change command and the determination result to the segment number change unit 130.

[0071] If the segment number change unit 130 of the information processing device 100B determines in the segment number change command reception determination process of step ST2310 that it has received a segment number change command (step ST2310 "Segment number change command received?" "YES"), it then executes a time series division result acquisition process (step ST2320 "Time series division result acquisition"). In this process, the segment number change unit 130 acquires the time series division result output by the time series division unit 101.

[0072] When the segment number change unit 130 of the information processing device 100B acquires the time-series division result, it then executes a change type determination process (step ST2330 "Determine whether to merge or divide segments based on the change command and the time-series division result"). In this process, the segment number change unit 130 decides whether to merge or divide the segments, which are the time-series division result, based on the time-series division result output by the time-series division unit 101 and the segment number change command received by the segment number change command receiving unit 112. In detail, for example, the time-series division unit 101 decides to switch between merging or dividing segments depending on the relationship between the number of segments N in the time-series division result and the command number of segments N' indicated in the segment number change command. A detailed example of the determination method will be described later.

[0073] If the segment number change unit 130 of the information processing device 100B determines in the change type determination process of step ST2330 that the process to merge segments is to be performed (step ST2330 "Deciding whether to merge or divide segments based on the change command and time series division result" "Merging"), it then executes a segment merging process (step ST2340 "Segment Merging"). The segment merging unit 131 of the segment number change unit 130 merges multiple segments that are subject to change and have similar characteristics among the segments in the time series division result into one segment to form the changed segment.

[0074] If the segment number change unit 130 of the information processing device 100B decides to split a segment in the process of step ST2330 (step ST2330 "Deciding whether to merge or split the cluster based on the change command and time-series splitting result" "Split"), it then executes a segment splitting process (step ST2350 "Segment splitting"). In the segment splitting process, the segment splitting unit 132 of the segment number change unit 130 splits the segment to be changed in such a way that it has the least impact on the segments other than the segment to be changed among the segments in the time-series splitting result.

[0075] The segment number changing unit 130 of the information processing device 100B then performs time-series segmentation result output processing (step ST2360 "Output time-series segmentation result"). This processing is the same as the time-series segmentation result output processing in step ST2040.

[0076] After the segment number change unit 130 of the information processing device 100B has executed the time-series division result output processing in step ST2360, or if it has determined in step ST2310 that it has not received a segment number change command (step ST2310 "Segment number change command received?" "NO"), it then executes a termination determination process (step ST2370 "Termination?"). This process is the same as the termination determination process in step ST1340. If the information processing device 100B determines to terminate in the termination determination process of step ST2370 (step ST2370 "Termination?" "YES"), it terminates the process shown in Figure 18 ("Termination").

[0077] A more detailed example of the process in steps ST2330, ST2340, and ST2350 shown in Figure 18 will be described below. Figure 19 is a flowchart showing a more detailed example of the process for changing the number of segments in the information processing device according to Embodiment 2 of this disclosure. In the information processing device 100B according to this embodiment, the segment number changing unit 130 executes, for example, the process shown in Figure 19 during the process shown in Figure 18. The segment number changing unit 130 of the information processing device 100B starts the process shown in Figure 19, following the process of step ST2320 (time-series division result acquisition process) shown in Figure 18. The segment number change unit 130 of the information processing device 100B then executes a command segment number acquisition process (step ST2410 "Acquire segment number N'"). In this process, the segment number change unit 130 acquires the segment number N' (command segment number N') indicated in the segment number change command output by the segment number change command receiving unit 112.

[0078] The segment number change unit 130 then executes a segment number change determination process (step ST2420 "N'!=N?"). In this process, the segment number change unit 130 determines whether the number of segments N in the time series division result matches the commanded number of segments N' indicated in the segment number change command. If the number of segments N in the time series division result does not match the commanded number of segments N' indicated in the segment number change command, the segment number change unit 130 determines to change the number of segments.

[0079] The segment number change unit 130 determines to change the segment number if the number of segments N in the time series division result does not match the number of segments N' indicated in the segment number change command (step ST2420 "N'!=N?" "NO"), and then executes a change type determination process (step ST2430 "N'>N?"). In this process, the segment number change unit 130 determines to execute a segment division process if the number of segments N' indicated in the segment number change command is greater than the number of segments N in the time series division result. The segment number change unit 130 determines to execute a segment merge process if the number of segments N' indicated in the segment number change command is less than the number of segments N in the time series division result.

[0080] If the segment number change unit 130 determines in the change type determination process that the command segment number N' indicated in the segment number change command is greater than the number of segments N in the time series division result (step ST2430 "N'>N?" "YES"), it then executes the segment division process (step ST2440 "Divide segments N'-N times"). In this process, the segment division unit 132 of the segment number change unit 130 executes the process of dividing the target segment from the segments in the time series division result, in a way that has the least impact on segments other than the target segment, N'-N times. If the segment number change unit 130 determines in the change type determination process that the command segment number N' indicated in the segment number change command is not greater than the number of segments N in the time series division result (step ST2430 "N'>N?" "NO"), it then executes a segment merging process (step ST2450 "Merge segments NN' times"). In this process, the segment merging unit 131 of the segment number change unit 130 executes a process NN' times to merge multiple segments that are to be changed and have similar characteristics, from among the segments in the time series division result, into one of the segments that will have the least impact on segments other than the segment to be changed, in order to create the changed segment. After the segment number changing unit 130 executes the segment splitting process in step ST2440 or the segment merging process in step ST2450, it then executes the process in step ST2360 (step ST2360 "Output time series splitting result").

[0081] Next, we will explain a detailed example of the process of changing the number of clusters in an information processing device. Figure 20 is a flowchart showing a more detailed example of the process of changing the number of clusters, which is one of the processes shown in Figure 17. The information processing device 100B according to this embodiment executes the process shown in Figure 20, for example, in the change command receiving unit 110 and the cluster number changing unit 120.

[0082] When the information processing device 100B starts the process shown in Figure 20 ("Start"), the cluster number change command receiving unit 111 of the change command receiving unit 110 in the information processing device 100B then executes a cluster number change command reception determination process (step ST2510 "Cluster number change command received?"). This process is the same as the cluster number change command reception determination process in step ST2060.

[0083] If the cluster number change unit 120 of the information processing device 100B determines in the cluster number change command reception determination process of step ST2510 that it has received a cluster number change command (step ST2510 "Cluster number change command received?" "YES"), it then executes a clustering result acquisition process (step ST2520 "Acquire clustering results"). In this process, the cluster number change unit 120 acquires the clustering results output by the clustering unit 102.

[0084] When the cluster number change unit 120 of the information processing device 100B obtains the clustering result, it then performs a change type determination process (step ST2530 "Determine whether to merge or split clusters based on the change command and clustering result"). In this process, the cluster number change unit 120 decides whether to merge or split the clusters, which are the clustering result, based on the clustering result output by the clustering unit 102 and the cluster number change command received by the change command reception unit 110. Specifically, for example, the cluster number change unit 120 decides to switch between merging or splitting clusters depending on the relationship between the number of clusters K in the clustering result and the command cluster number K' indicated in the cluster number change command. A detailed example of the determination method will be described later.

[0085] If the cluster number change unit 120 of the information processing device 100B determines in the change type determination process of step ST2530 that it is to merge clusters (step ST2530 "Deciding whether to merge or split clusters based on the change command and clustering results" "Merge"), it then executes the cluster merging process (step ST2540 "Cluster Merging"). The cluster merging unit 121 of the cluster number change unit 120 merges multiple clusters that are subject to change and have similar characteristics among the clusters in the clustering results into one cluster to form the changed cluster.

[0086] If the cluster number change unit 120 of the information processing device 100B decides to split a cluster in the process of step ST2530 (step ST2530 "Deciding whether to merge or split clusters based on change command and clustering results" "Split"), it then executes a cluster splitting process (step ST2550 "Cluster splitting"). In the cluster splitting process, the cluster splitting unit 122 of the cluster number change unit 120 splits the cluster to be changed in a way that has the least impact on clusters other than the cluster to be changed among the clusters in the clustering results.

[0087] The cluster number changing unit 120 of the information processing device 100B then executes the clustering result output process (step ST2560 "Output clustering results"). This process is the same as the clustering result output process in step ST2080.

[0088] After executing the process in step ST2560, or if the cluster number change unit 120 of the information processing device 100B determines in step ST2510 that it did not receive a cluster number change command (step ST2510 "Cluster number change command received?" "NO"), it then executes a termination determination process (step ST2570 "Termination?"). This process is the same as the termination determination process in step ST2090.

[0089] If the information processing device 100B determines in the termination determination process of step ST2570 that it is not terminating (step ST2570 "Termination?" "NO"), it returns to the process of step ST2510. If the information processing device 100B determines in the termination determination process of step ST2570 that it should terminate (step ST2570 "Termination?" "YES"), it terminates the process shown in Figure 20 ("Termination").

[0090] Here, we will describe a more detailed example of the process in steps ST2530, ST2540, and ST2550 shown in Figure 20. Figure 21 is a flowchart showing a more detailed example of the process for changing the number of clusters in the information processing device according to Embodiment 2 of this disclosure. In the information processing device 100B according to this embodiment, the cluster number changing unit 120 executes, for example, the process shown in Figure 21 during the process shown in Figure 20. After the cluster number changing unit 120 executes the process of step ST2520 shown in Figure 20, it then starts the process shown in Figure 21.

[0091] The cluster number change unit 120 of the information processing device 100B then executes a command cluster number acquisition process (step ST2610 "Acquire cluster number K'"). In this process, the cluster number change unit 120 acquires the cluster number K' (command cluster number K') indicated in the cluster number change command output by the cluster number change command receiving unit.

[0092] The cluster number change unit 120 then executes a cluster number change determination process (step ST2620 "K'!=K?"). In this process, the cluster number change unit 120 determines whether the cluster number K of the clustering result matches the command cluster number K' indicated in the cluster number change command. If the cluster number K of the clustering result does not match the command cluster number K' indicated in the cluster number change command, the cluster number change unit 120 determines to change the cluster number. The cluster number change unit 120 determines that it will not change the cluster number if the cluster number K of the clustering result does not match the command cluster number K' indicated in the cluster number change command (step ST2620 "K'!=K?" "YES"), and proceeds to the clustering result output processing in step ST2560, as shown in Figure 20.

[0093] The cluster number change unit 120 determines to change the cluster number if the cluster number K of the clustering result does not match the command cluster number K' indicated in the cluster number change command (step ST2620 "K'!=K?" "NO"), and then executes a change type determination process (step ST2630 "K'>K?"). In this process, the cluster number change unit 120 determines to execute a cluster splitting process if the command cluster number K' indicated in the cluster number change command is greater than the cluster number K of the clustering result. The cluster number change unit 120 determines to execute a cluster merging process if the command cluster number K' indicated in the cluster number change command is less than the cluster number K of the clustering result.

[0094] The cluster number change unit 120, if the command cluster number K' indicated in the cluster number change command is greater than the cluster number K in the clustering result (step ST2630 "K'>K?" "YES"), then executes a cluster partitioning process (step ST2640 "Partition clusters K'-K times"). In this process, the cluster partitioning unit 122 of the cluster number change unit 120 executes a process K'-K times to partition the cluster to be changed from among the clusters in the clustering result in such a way that it has the least impact on clusters other than the cluster to be changed.

[0095] The cluster number change unit 120, if the command cluster number K' indicated in the cluster number change command is smaller than the cluster number K in the clustering result (step ST2630 "K'>K?" "NO"), then executes a cluster merging process (step ST2650 "Merge clusters KK' times"). In this process, the cluster merging unit 121 of the cluster number change unit 120 executes a process KK' times to merge multiple clusters that are the target of modification, which have the least impact on clusters other than the target cluster in the clustering result and have similar characteristics, into one cluster to create the modified cluster.

[0096] When the information processing device 100A according to this embodiment 1 executes the cluster partitioning process in step ST1540 or the cluster merging process in step ST1550, it then terminates the series of processes shown in Figure 9 ("termination") and proceeds to the clustering result output process shown in Figure 20 (step ST2560 "output clustering results").

[0097] According to the configuration of this embodiment, users can modify not only undesirable clustering results but also undesirable time-series segmentation, enabling more consistent and flexible time-series segmentation. Furthermore, by merging segments that have been created due to time-series-specific chattering, an improvement in clustering accuracy can be expected.

[0098] This embodiment further illustrates an example of a configuration that can be expressed as follows. [5] The aforementioned change command receiving unit further receives a segment number change command indicating the commanded number of segments, When the segment number change command is received, the segment number change unit determines and executes the merging or splitting of the segments, which are the result of the time series data being divided into segments, based on the time series division result and the segment number change command, and outputs the time series division result after the segment number has been changed. An information processing device according to any one of [1][2][3][4], further comprising the above. This disclosure further provides an information processing device that enables maintaining clustering consistency when modifying clustering results. Furthermore, by applying the above-mentioned information processing device to a time-series segmentation device, it is possible to provide a time-series segmentation device that, similarly to the above, enables the maintenance of clustering consistency when modifying clustering results. Furthermore, by applying the above configuration to a system including the above information processing device, the above information processing method, or the above program, etc., this disclosure will produce the same effects as described above.

[0099] This embodiment further illustrates an example of a configuration that can be expressed as follows. [6] The segment number changing unit is, A segment division unit divides the segments to be changed from the aforementioned time-series division results and maintains the segments other than those to be changed. A segment merging unit merges the segments to be changed from the clustering results and maintains the segments that are not to be changed, Equipped with, The information processing apparatus according to [5], characterized in that This disclosure further provides an information processing device that enables maintaining clustering consistency when modifying clustering results. Furthermore, by applying the above-mentioned information processing device to a time-series segmentation device, it is possible to provide a time-series segmentation device that enables the maintenance of clustering consistency when modifying clustering results, similar to the above. Furthermore, by applying the above configuration to a system including the above information processing device, the above information processing method, or the above program, etc., this disclosure will produce the same effects as described above.

[0100] Embodiment 3. In Embodiment 1 described above, an example configuration was shown in which the information processing device automatically extracts the clusters to be modified and performs cluster merging or splitting. Furthermore, in Embodiment 2 described above, an example configuration was shown in which the information processing device automatically extracts the segments to be modified and performs segment merging or splitting. Embodiment 3 describes an example of a configuration in which the target to be changed can be selected and specified from outside the information processing device. In Embodiment 3, components related to Embodiment 3 that are the same as those related to Embodiment 1 or Embodiment 2 already described are denoted by the same reference numerals or similar numerals, and redundant explanations are omitted as appropriate.

[0101] Next, an example of the configuration of the information processing device according to Embodiment 3 of this disclosure will be described. The information processing device (time-series segmentation device) 100 (100C) and the information processing system (time-series segmentation system) 1 (1C) including the information processing device according to this third embodiment differ from the configuration of the embodiment already described in that, when changing the clustering result, it accepts the selection and specification of clusters, segments, or clusters and segments to be changed, and has the function to merge or split the selected targets to be changed.

[0102] Figure 22 is a diagram showing a first configuration example of an information processing device and an information processing system including the information processing device according to Embodiment 3 of the present disclosure. The information processing system (time-series segmentation system) 1(1C) shown in Figure 22 is composed of an input / output device 30 and an information processing device (time-series segmentation device) 100(100C). The input / output device 30 is configured to include an input source device 10 and an output destination device 20. The output device 20 outputs an image representing the processing result output by the information processing device 100 (100C), similar to the functions already described. (For example, an image representing the clustering result, the time series partitioning result, or both.) In addition to the functions already described, the input source device 10 has the function of allowing the information processing device 100 (100C) to input information that allows the user to select and specify the target for modification (cluster or segment) based on the user's operation regarding the processing result (for example, an image representing the clustering result, the time series partitioning result, or both).

[0103] The information processing device 100 (100C) receives selected clusters from the clustering results, which are selected and specified by a user or other external party, and changes the number of clusters by merging or splitting the selected clusters. The information processing device 100 (100C) shown in Figure 22 is composed of a time-series division unit 101, a clustering unit 102, a change command reception unit 110, a cluster number change unit 120, and a selection reception unit 140. The change command receiving unit 110 shown in Figure 22 is configured to include a cluster number change command receiving unit 111. The cluster number changing unit 120 shown in Figure 22 is composed of a cluster merging unit 121 and a cluster splitting unit 122.

[0104] The selection reception unit 140 accepts selection specifications for items to be changed when modifying the clustering results. The selection reception unit 140 accepts selection specifications indicating items to be changed, for example, selected by the user via the input source device 10. The selection reception unit 140 shown in Figure 22 is configured to include a selection cluster reception unit 141. The selected cluster reception unit 141 receives a selected cluster, which is a cluster selected from among the clusters in the clustering result. In detail, for example, the selected cluster receiving unit 141 receives the selected cluster, which is the cluster to be modified selected from the clusters in the clustering result, via the input source device 10 when the user operates the input source device 10 to select a cluster to be modified from the clusters in the clustering result output by the output destination device 20. The cluster number change unit 120 performs cluster merging or splitting using the selected clusters received by the selected cluster reception unit 141. In detail, for example, the cluster number change unit 120, when changing the number of clusters, accepts selected clusters corresponding to merging or splitting, and performs merging or splitting of the accepted selected clusters.

[0105] Figure 23 is a diagram showing a second configuration example of an information processing device and an information processing system including the information processing device according to Embodiment 3 of the present disclosure. The information processing device 100 (100C) shown in Figure 23 receives selected segments from the time-series segmentation results, which are selected and specified by a user or other external party, and changes the number of segments by merging or splitting the selected segments. Furthermore, the information processing device 100 (100C) receives selected clusters from the clustering results, which are selected and specified by a user or other external party, and changes the number of clusters by merging or splitting the selected clusters. The information processing device 100 (100C) shown in Figure 23 is composed of a time-series division unit 101, a clustering unit 102, a change command reception unit 110, a cluster number change unit 120, a segment number change unit 130, and a selection reception unit 140. The change command receiving unit 110 shown in Figure 23 is composed of a cluster number change command receiving unit 111 and a segment number change command receiving unit 112. The cluster number changing unit 120 shown in Figure 23 is composed of a cluster merging unit 121 and a cluster splitting unit 122. The segment number changing unit 130 shown in Figure 23 is composed of a segment merging unit 131 and a segment splitting unit 132.

[0106] The selection receiving unit 140 receives a selection command to change the target when modifying the clustering result. The selection receiving unit 140 receives a selection command indicating the segment or cluster to be modified, which has been selected by the user, for example, via the input source device 10. The selection reception unit 140 shown in Figure 23 is configured to include a selection cluster reception unit 141 and a selection segment reception unit 142. In other words, the selection reception unit 140 includes a selection segment reception unit 142 in addition to the selection cluster reception unit 141.

[0107] The selected segment receiving unit 142 receives selected segments, which are segments selected from among the segments into which the time-series data has been divided. The segment number change unit 130 performs segment merging or splitting using a change command (segment number change command) or the selected segments. Specifically, the segment number change unit 130 obtains the change command (segment number change command) received by the segment number change command receiving unit 112 of the change command receiving unit 110, the time series splitting result by the time series splitting unit 101, and the selected segments received by the selected segment receiving unit 142 of the selection receiving unit 140. Based on the change command (segment number change command), the selected segments, and the time series splitting result, the segment number change unit 130 decides whether to merge or split the segments. For example, if the change command includes a change type of merge or split, the segment number change unit 130 decides whether to merge or split the segments based on the change command. For example, if the number of selected segments received is multiple, the segment number change unit 130 decides to perform segment merging processing, and if the number of selected segments received is one, it decides to perform segment splitting processing. The segment merging unit 131 of the segment number change unit 130 merges the selected segments, which are clusters selected as targets for change from the clustering results, while maintaining the segments other than those targeted for change. If the segment number change unit 130 switches to the merging process, it executes the process. The segment division unit 132 of the segment number change unit 130 divides the selected segments, which are clusters selected as targets for change from the time-series division results, and maintains the segments other than those targeted for change. The segment division unit 132 executes processing when the segment number change unit 130 switches to division processing.

[0108] The selected cluster reception unit 141 accepts selected clusters, which are clusters selected from among the clusters in the clustering results. The cluster number change unit 120 performs cluster merging or splitting using a change command (cluster number change command) or the selected cluster. Specifically, the cluster number change unit 120 obtains the change command (cluster number change command) received by the cluster number change command receiving unit 111 of the change command receiving unit 110, the clustering results from the clustering unit 102, and the selected cluster received by the selected cluster receiving unit 141 of the selection receiving unit 140. The cluster number change unit 120 decides whether to merge or split the cluster based on the change command (cluster number change command), the selected cluster, and the time-series splitting results. For example, if the change command includes a change type of merge or split, the cluster number change unit 120 decides whether to merge or split the cluster based on the change command. For example, if the number of selected clusters received is multiple, the cluster number change unit 120 decides to perform cluster merging processing, and if the number of selected clusters received is one, it decides to perform cluster splitting processing. The cluster merging unit 121 of the cluster number change unit 120 merges the selected clusters, which are the clusters selected for modification from the clustering results, while maintaining the clusters that are not targeted for modification. The cluster merging unit 121 executes the process when the cluster number change unit 120 switches to the merging process. The cluster division unit 122 of the cluster number change unit 120 divides the selected clusters, which are the clusters selected for modification from among the clusters of the clustering result, and maintains the clusters that are not to be modified. The cluster division unit 122 executes processing when the cluster number change unit 120 switches to the division process.

[0109] Next, an example of processing by the information processing device according to Embodiment 3 of this disclosure will be described. Figure 24 is a flowchart showing an example of a process for changing the number of segments in an information processing device according to Embodiment 3 of this disclosure. The process shown in Figure 24 is the process obtained by adding the process according to this embodiment to the process shown in Figure 18. The information processing device 100C according to this embodiment executes the processing shown in Figure 24 using a change command receiving unit 110 (a segment number change command receiving unit 112 in the change command receiving unit 110), a selection receiving unit 140 (a selected segment receiving unit 142 in the selection receiving unit 140), and a segment number change unit 130. The information processing device 100C performs the time-series division process shown in step ST2010 in Figure 17, for example, and then starts the process shown in Figure 24 ("start"). When the information processing device 100C starts the process shown in Figure 24, it then executes a segment number change command reception determination process (step ST3110 "Segment number change command received?"). In the segment number change command reception determination process, the segment number change command reception unit 112 in the change command reception unit 110 of the information processing device 100C checks whether a segment number change command has been received and makes a determination, similar to the process in step ST2020 already described. If it determines that a segment number change command has been received (step ST3110 "Segment number change command received?" "YES"), it outputs the segment number change command and the determination result to the segment number change unit 130. If the segment number change command reception unit 112 determines that it has not received a segment number change command (step ST3110 "Cluster number change command received?" "NO"), it proceeds to the termination determination process in step ST3190. The information processing device 100C then performs a time-series segmentation result acquisition process (step ST3120 "Time-series segmentation result acquisition"). In the time-series segmentation result acquisition process, the segment number change unit 130 of the information processing device 100C acquires the time-series segmentation result output by the time-series segmentation unit 101. The information processing device 100C then executes a selection segment acceptance determination process (step ST3130 "Selection segment accepted?"). In the selection segment acceptance determination process, the selection segment acceptance unit 142 of the selection acceptance unit 140 of the information processing device 100C determines that a selection segment has been accepted if a segment is selected from the segments of the time-series division result via the input source device 10 based on user operation within a predetermined time period, and notifies the segment number change unit 130 of the determination result, which includes the determination result and information on the selected segment. If no segment is selected from the segments of the time-series division result via the input source device 10 based on user operation within a predetermined time period, the selection segment acceptance unit 142 determines that a selection segment has not been accepted, and notifies the segment number change unit 130 of the determination result. If the information processing device 100C determines that it did not accept the selected segment (step ST3130 "Selected segment accepted?" "NO"), it proceeds to step ST2330 in Figure 18, which was previously described. After executing the processing in step ST2330 and the processing in step ST2340 or step ST2350, it returns to the termination determination process in step ST3190. If the information processing device 100C determines that it has received a selected segment (step ST3130 "Selected segment received?" "YES"), it then executes the selected segment acquisition process (step ST3140 "Selected segment acquired"). In the selected segment acquisition process, the segment number change unit 130 of the information processing device 100C acquires information relating to the selected segment from the segments of the time-series division result. The information processing device 100C then performs a change type determination process (step ST3150 “Determine whether to merge or split segments based on the change command, selected segments, and time-series splitting results”). In the change type determination process, the segment number change unit 130 of the information processing device 100C decides whether to merge or split segments based on the change command (segment number change command), selected segments, and time-series splitting results. For example, if the change command includes a change type of merge or split, the segment number change unit 130 decides whether to merge or split segments based on the change command. For example, if the number of selected segments received is multiple, the segment number change unit 130 decides to perform a segment merge process, and if the number of selected segments received is one, it decides to perform a segment split process. If the segment number change unit 130 of the information processing device 100C decides to perform segment merging in the change type determination process of step ST3150 (step ST3150 "merging"), it then performs segment merging processing (step ST3160 "segment merging"). In the segment merging processing, the cluster merging unit 121 of the segment number change unit 130 of the information processing device 100C merges multiple selected segments. If the information processing device 100C decides to perform segment division in the change type determination process of step ST3150 (step ST3150 "division"), it then performs segment division processing (step ST3170 "segment division"). In the segment division processing, the segment division unit 132 of the segment number change unit 130 of the information processing device 100C divides the selected segment. After the information processing device 100C executes the segment merging process in step ST3160 or the segment splitting process in step ST3170, it then executes the time-series splitting result output process (step ST3180 "Output time-series splitting result"). In the time-series splitting result output process, the segment number change unit 130 of the information processing device 100C outputs the time-series splitting result, including information related to the segments, to the clustering unit 102. The segment number change unit 130 also outputs the time-series splitting result, including information related to the segments, to the output destination device 20. After the information processing device 100C has executed the time-series division result output processing in step ST3180, or the processing in steps ST2330 to ST2340 or ST2350, it then executes a termination determination process (step ST3190 "Termination?"). In the termination determination process, the control unit (not shown) of the information processing device 100C determines to terminate the process, for example, if it determines that the input of time-series data has finished. Alternatively, it determines to terminate the process, for example, if it receives a termination command from an external source. If the information processing device 100C determines in the termination determination process of step ST3190 that it is not terminating (step ST3190 "Termination?" "NO"), it returns to the process of step ST3110. If the information processing device 100C determines to terminate in the termination determination process of step ST3190 (step ST3190 "Termination?" "YES"), it terminates the process shown in Figure 24 ("Termination").

[0110] Figure 25 is a flowchart showing an example of a process for changing the number of clusters in an information processing device according to Embodiment 3 of this disclosure. The process shown in Figure 25 is the process obtained by adding the process according to this embodiment to the process shown in Figure 20. The information processing device 100C according to this embodiment executes the process shown in Figure 25, for example, using the change command receiving unit 110 (cluster number change command receiving unit 111 in the change command receiving unit 110), the selection receiving unit 140 (selected cluster receiving unit 141 in the selection receiving unit 140), and the cluster number change unit 120. The information processing device 100C starts the process shown in Figure 25 ("Start") after performing the clustering process shown in step ST2050, for example, as shown in Figure 17. When the information processing device 100C starts the process shown in Figure 25, it then executes the cluster number change command reception determination process (step ST3210 "Cluster number change command received?"). In the cluster number change command reception determination process, the cluster number change command reception unit 111 of the change command reception unit 110 of the information processing device 100C checks whether a cluster number change command has been received and makes a determination, similar to the process in step ST2060 already described. If it determines that a cluster number change command has been received (step ST3210 "Cluster number change command received?" "YES"), it outputs the cluster number change command and the determination result to the cluster number change unit 120. If the cluster number change command reception unit 111 determines that a cluster number change command has not been received (step ST3210 "Cluster number change command received?" "NO"), it proceeds to the termination determination process in step ST3290. If the information processing device 100C determines in step ST3210 that it has received a command to change the number of clusters (step ST3210 "Command to change the number of clusters received?" "YES"), it then executes the clustering result acquisition process (step ST3220 "Acquire clustering results"). In the clustering result acquisition process, the cluster number change unit 120 of the information processing device 100C acquires the clustering results output by the clustering unit 102. The information processing device 100C then executes a selection cluster acceptance determination process (step ST3230 "Selection cluster accepted?"). In the selection cluster acceptance determination process, the selection cluster acceptance unit 141 of the selection acceptance unit 140 of the information processing device 100C determines that a selection cluster has been accepted if a cluster is selected from the clusters of the clustering result via the input source device 10 based on user operation within a predetermined time period, and notifies the cluster count change unit 120 of the determination result, which includes the determination result and information on the selected cluster. If no cluster is selected from the clusters of the clustering result via the input source device 10 based on user operation within a predetermined time period, the selection cluster acceptance unit 141 determines that a selection cluster has not been accepted, and notifies the cluster count change unit 120 of the determination result. If the information processing device 100C determines that it did not accept the selected cluster (step ST3230 "Selected cluster accepted?" "NO"), it proceeds to step ST2530 as shown in Figure 18, which was previously explained. After executing the processing in step ST2530 and the processing in step ST2540 or step ST2550, it returns to the termination determination process in step ST3290. If the information processing device 100C determines that it has received the selected cluster (step ST3230 "Selected cluster received?" "YES"), it then executes the selected cluster acquisition process (step ST3240 "Selected cluster acquisition"). In the selected cluster acquisition process, the cluster number change unit 120 of the information processing device 100C acquires information relating to the selected cluster from among the clusters of the clustering result. The information processing device 100C then performs a change type determination process (step ST3250 “Determine whether to merge or split clusters based on the change command and clustering results”). In the change type determination process, the cluster number change unit 120 of the information processing device 100C decides whether to merge or split segments based on the change command (cluster number change command), the selected clusters, and the time-series splitting results. For example, if the change command includes a change type of merge or split, the cluster number change unit 120 decides whether to merge or split clusters based on the change command. For example, if the number of selected clusters received is multiple, the cluster number change unit 120 decides to perform a cluster merge process, and if the number of selected clusters received is one, it decides to perform a cluster split process. If the information processing device 100C decides to perform cluster merging in the change type determination process of step ST3250 (step ST3250 "Merge"), it then performs the cluster merging process (step ST3260 "Cluster Merging"). The cluster merging unit 121 in the cluster number change unit 120 of the information processing device 100C merges multiple selected clusters. If the information processing device 100C decides to perform cluster division in the change type determination process of step ST3250 (step ST3250 "division"), it then performs cluster division processing (step ST3270 "cluster division"). In the cluster division processing, the cluster division unit 122 of the cluster number change unit 120 of the information processing device 100C divides the selected cluster. After the information processing device 100C executes the cluster merging process in step ST3260 or the cluster splitting process in step ST3270, it then executes the clustering result output process (step ST3280 "Output clustering results"). In the clustering result output process, the cluster number changing unit 120 of the information processing device 100C outputs the clustering results, including information related to the clusters, to the output destination device 20. The cluster number changing unit 120 of the information processing device 100C, after executing the time-series division result output processing in step ST3280, or the processing in steps ST2530 and ST2540 or ST2550, then executes a termination determination process (step ST3290 "Termination?"). In the termination determination process, the control unit (not shown) of the information processing device 100C determines to terminate the process, for example, if it determines that the input of time-series data has finished. Alternatively, for example, it determines to terminate the process if it receives a termination command from an external source. If the information processing device 100C determines in the termination determination process of step ST3290 that it is not terminating (step ST3290 "Termination?" "NO"), it returns to the process of step ST3210. If the information processing device 100C determines to terminate in the termination determination process of step ST3290 (step ST3290 "Termination?" "YES"), it terminates the process shown in Figure 25 ("Termination").

[0111] Here, a detailed second example of a process for changing the number of segments or the number of clusters according to this embodiment will be described. First, let's explain a second detailed example of the process of changing the number of segments. Figure 26 is a flowchart showing a detailed example of the process for changing the number of segments in the information processing device according to Embodiment 3 of this disclosure. In the information processing device 100C according to this embodiment, the segment number changing unit 130 executes, for example, the process shown in Figure 26 during the process shown in Figure 17. The segment number changing unit 130 starts the process shown in Figure 26 during the segment number changing process (step ST2030) shown in Figure 17, for example. The segment number change unit 130 then executes a segment number change command information acquisition process (step ST3310 "Acquire segment number change command"). The segment number change command information includes the change type, such as merging or splitting, which is input by the user via the input source device 10. The segment number change unit 130 then performs a change method determination process (step ST3320 "Splitting process?"). In the change method determination process, the segment number change unit 130 determines whether it is a splitting process based on the segment number change command information. If the segment number change unit 130 determines in the change method determination process of step ST3320 that it is a splitting process (step ST3320 "Splitting process?" "YES"), it then executes the process of acquiring selected segments for splitting (step ST3330 "Acquire selected segments to be splitted"). In the selected segment acquisition process, the segment number change unit 130 acquires selected segment information for splitting from the selected segment receiving unit 142, which indicates the segments to be splitted that have been selected and specified by the user via the output destination device 20. Following the processing in step ST3330, the segment number change unit 130 executes the segment splitting process (step ST3340 "execution of segment splitting"). In the segment splitting process of step ST3340, the segment number change unit 130 splits the selected segments for splitting. The splitting method has already been described in the embodiment described above, so it will not be described here. If the segment number change unit 130 determines in the change method determination process of step ST3320 that it is not a splitting process (step ST3320 "Splitting process?" "NO"), it then executes the process of acquiring selected segments for merging (step ST3350 "Acquire selected segments to be merged"). In the process of acquiring selected segments for merging in step ST3350, the segment number change unit 130 acquires selected segment information for merging, which indicates the segments to be merged selected by the user via the output destination device 20, from the selected segment receiving unit 142. The segment number change unit 130 executes a segment merging process (step ST3360 "execution of segment merging") following the process of acquiring selected segments for merging in step ST3350. In the segment merging process, the segment number change unit 130 merges the selected segments for merging. The merging method is the same as in the embodiment already described, so its description is omitted here. When the segment number changing unit 130 executes the process of step ST3340 or step ST3360, it terminates the series of processes shown in Figure 26 and proceeds to, for example, step ST2040 in Figure 17. The segment number change unit 130 outputs the time-series segment division results after processing by the segment division process to the output destination device 20. The segment number change unit 130 also outputs to the clustering unit 102. The segment number change unit 130 outputs the time-series division results after processing by the merger process to the output destination device 20. The segment number change unit 130 also outputs the time-series division results after processing by the merger process to the clustering unit 102.

[0112] Next, we will explain a second detailed example of the process of changing the number of clusters. Figure 27 is a flowchart showing a detailed example of the process for changing the number of clusters in the information processing device according to Embodiment 3 of this disclosure. In the information processing device 100C according to this embodiment, the cluster number changing unit 120 executes, for example, the process shown in Figure 27 during the process shown in Figure 17. The cluster number change unit 120 starts the process shown in Figure 27 during the cluster number change process (step ST2070) shown in Figure 17. The cluster number change unit 120 then executes the process of acquiring cluster number change command information (step ST3410 "Acquire cluster number change command"). The cluster number change command information includes the type of change, such as merging or splitting, which is input by the user via the input source device 10. The cluster number change unit 120 then performs a change method determination process (step ST3420 "Splitting process?"). In the change method determination process, the cluster number change unit 120 determines whether it is a splitting process based on the cluster number change command information. If the cluster number change unit 120 determines in the change method determination process of step ST3420 that it is a splitting process (step ST3420 "Splitting process?" "YES"), it then executes the process of acquiring selected clusters for splitting (step ST3430 "Acquire selected clusters to be splitted"). In the process of step ST3430, the cluster number change unit 120 acquires selected cluster information for splitting from the selected cluster reception unit 141, which indicates the clusters to be splitted that have been selected and specified by the user via the output destination device 20. The cluster number change unit 120 executes the cluster partitioning process following the processing in step ST3430 (step ST3440 "Execute cluster partitioning"). In the segment partitioning process of step ST3340, the segment number change unit 130 partitions the selected segments for partitioning. The partitioning method is the same as in the embodiment already described, so the explanation is omitted here. If the cluster number change unit 120 determines in the change method determination process of step ST3420 that it is not a splitting process (step ST3420 "Splitting process?" "NO"), it then executes the process of acquiring selected clusters for merging (step ST3450 "Acquire selected clusters to be merged"). In the process of acquiring selected clusters for merging, the cluster number change unit 120 acquires selected cluster information for merging from the selected cluster reception unit 141, which indicates the clusters to be merged that have been selected by the user via the output destination device 20. The cluster number change unit 120 executes the cluster merging process following the processing in step ST3450 (step ST3460 "Execute cluster merging"). In the cluster merging process of step ST3460, the cluster number change unit 120 merges the selected clusters for merging. The merging method is the same as in the embodiment already described, so its description is omitted here. When the cluster number changing unit 120 executes the process of step ST3340 or step ST3360, it terminates the series of processes shown in Figure 27 and proceeds to, for example, step ST2080 in Figure 17. The cluster number change unit 120 outputs the clustering results after processing by the partitioning process to the output destination device 20. The cluster number change unit 120 also outputs the clustering results after processing by the partitioning process to the clustering unit 102. The cluster number change unit 120 outputs the clustering results after processing by the merger process to the output destination device 20. The cluster number change unit 120 also outputs the clustering results after processing by the merger process to the clustering unit 102.

[0113] According to the configuration described above, when changing the number of clusters, users can select and merge or split segments, clusters, or both based on their own knowledge. This allows for manual correction even in cases where, for example, "domain knowledge suggests they should be grouped into the same cluster, but the algorithm results in them being in separate clusters," thereby improving the user experience.

[0114] This embodiment further illustrates an example of a configuration that can be expressed as follows. [3] A selected cluster receiving unit receives a selected cluster, which is a cluster selected from among the clusters in the clustering results. Equipped with, The cluster number changing unit performs cluster merging or splitting using the selected clusters. An information processing device according to either [1] or [2], characterized in that This disclosure further provides an information processing device that allows users to select and merge or split clusters based on their knowledge, and that maintains clustering consistency when modifying clustering results. Furthermore, by applying the above-mentioned information processing device to a time-series segmentation device, it is possible to provide a time-series segmentation device that enables the maintenance of clustering consistency when modifying clustering results, similar to the above. Furthermore, by applying the above configuration to a system including the above information processing device, the above information processing method, or the above program, etc., this disclosure will produce the same effects as described above.

[0115] This embodiment further illustrates an example of a configuration that can be expressed as follows. [7] A selection segment receiving unit that receives a selection segment, which is a segment selected from the segments into which the time-series data has been divided. Equipped with, The segment number changing unit performs segment merging or splitting using the selected segments. An information processing apparatus according to any one of [5] or [6], characterized in that This disclosure further provides an information processing device that can leverage user knowledge to select and merge or split segments, and that can maintain clustering consistency when modifying clustering results. Furthermore, by applying the above-mentioned information processing device to a time-series segmentation device, it is possible to provide a time-series segmentation device that, similarly to the above, enables the maintenance of clustering consistency when modifying clustering results. Furthermore, by applying the above configuration to a system including the above information processing device, the above information processing method, or the above program, etc., this disclosure will produce the same effects as described above.

[0116] Embodiment 4. In the above-described embodiment 3, a configuration was described in which it is possible to select and specify the target for changing the number of clusters. Embodiment 4 further describes a configuration that enables assistance when selecting and specifying the target to be changed. In Embodiment 4, components of Embodiment 4 that are the same as those of Embodiment 1, Embodiment 2, or Embodiment 3 already described are denoted by the same reference numerals or similar numerals, and redundant explanations are omitted as appropriate.

[0117] Next, an example of the configuration of the information processing device according to Embodiment 4 of this disclosure will be described. The information processing device (time-series segmentation device) 100 (100D) and the information processing system (time-series segmentation system) 1 (1D) including the information processing device according to this fourth embodiment differ from the configuration according to the third embodiment already described in that it has a function to assist the user in selecting the cluster, segment, or cluster and segment to be changed.

[0118] Figure 28 is a diagram showing a first configuration example of an information processing device and an information processing system including the information processing device according to Embodiment 4 of the present disclosure. The system consists of the information processing system (time-series segmentation system) 1(1D), input / output device 30, and information processing device (time-series segmentation device) 100(100D) shown in Figure 28. The input / output device 30 is configured to include an input source device 10 and an output destination device 20. The output device 20 outputs an image representing the processing results output by the information processing device 100 (100D), similar to the functions already described. (For example, an image representing the clustering results, time-series segmentation results, or both.) The output device 20 also outputs a selection support screen that indicates selection candidates when prompting the user to select the items to be changed (segmentation targets, merger targets). In addition to the functions already described, the input source device 10 has the function of allowing the information processing device 100 (100C) to input information that allows the user to select and specify the target for modification (cluster or segment) based on the user's operation regarding the processing result (for example, an image representing the clustering result, the time series partitioning result, or both).

[0119] When the information processing device 100 (100D) receives a selected cluster from the clustering results, which has been selected and specified by a user or other external party, it instructs the user on a suitable candidate cluster for modification. The information processing device 100 (100D) shown in Figure 28 is composed of a time-series division unit 101, a clustering unit 102, a change command reception unit 110, a cluster number change unit 120, a selection reception unit 140, and a selection support unit 150. The change command receiving unit 110 shown in Figure 28 is configured to include a cluster number change command receiving unit 111. The cluster number changing unit 120 is configured to include a cluster merging unit 121 and a cluster splitting unit 122, similar to the embodiment described earlier. The selection reception unit 140 shown in Figure 28 is configured to include a selection cluster reception unit 141.

[0120] The selection support unit 150 indicates candidate selections before accepting the selection of the cluster, segment, or both the cluster and segment to be changed. The selection support unit 150 shown in Figure 28 includes a cluster selection support unit 151. The cluster selection support unit 151 extracts candidate clusters based on the characteristics of each cluster in the clustering result (feature distance, distance based on features, distance in feature space) and outputs an image indicating the candidate cluster.

[0121] Figure 29 is a diagram showing a second configuration example of an information processing device and an information processing system including the information processing device according to Embodiment 4 of the present disclosure. The information processing device 100 (100D) shown in Figure 29, when it receives a selected segment from the time-series segmentation results that has been selected and specified by a user or other external party, instructs the user on a suitable candidate segment for modification. Furthermore, when the information processing device 100 (100D) receives a selected cluster from the clustering results that has been selected and specified by a user or the like from outside the device, it instructs the user on a suitable candidate cluster for modification. The information processing device 100 (100D) shown in Figure 29 is composed of a time-series division unit 101, a clustering unit 102, a change command reception unit 110, a cluster number change unit 120, a selection reception unit 140, and a selection support unit 150. The change command receiving unit 110 shown in Figure 29 is configured to include a cluster number change command receiving unit 111 and a segment number change command receiving unit 112, similar to the change command receiving unit 110 of the embodiment described earlier. The cluster number changing unit 120 is configured to include a cluster merging unit 121 and a cluster splitting unit 122, similar to the cluster number changing unit 120 in the embodiment described above. The segment number changing unit 130 is configured to include a segment merging unit 131 and a segment splitting unit 132, similar to the segment number changing unit 130 of the embodiment described above. The selection reception unit 140 shown in Figure 29 is configured to include a selection cluster reception unit 141, similar to the embodiment 3 described earlier.

[0122] The selection support unit 150 assists the user in selecting segments when the information processing device 100 (100D) receives a command to change the number of segments in the time-series division result. When the selection support unit 150 receives a command to change the number of segments from the information processing device 100 (100D), it indicates candidate segments for selection. Furthermore, the selection support unit 150 assists the user in selecting a cluster when the information processing device 100 (100D) receives a command to change the number of clusters in the clustering results. When the selection support unit 150 receives a command to change the number of clusters from the information processing device 100 (100D), it indicates candidate clusters for selection. The selection support unit 150 shown in Figure 29 is composed of a cluster selection support unit 151 and a segment selection support unit 152. In other words, the selection support unit 150 includes a segment selection support unit 152 in addition to the cluster selection support unit 151.

[0123] The segment selection support unit 152 extracts candidate segments based on the characteristics of each segment in the time-series segmentation result and outputs an image indicating the candidate segment. Features can be represented, for example, by a distance scale of features between segments (also called feature distance or distance in feature space) derived from clustering or segmentation algorithms. The segment selection support unit 152 outputs an image indicating the candidate segment by applying visual effects such as color, pattern, or blinking to segments with a small distance scale. For example, when one segment is selected by the user, the segment selection support unit 152 may indicate segments with a small feature distance scale as candidate segments based on the selected segment, or it may indicate multiple segments with a small distance scale as candidate segments when no segment is selected.

[0124] The cluster selection support unit 151 extracts candidate clusters based on the characteristics of each cluster in the clustering results and outputs an image indicating the candidate clusters. Features can be represented, for example, by a distance scale (feature distance, or distance in feature space) between features derived from clustering or segmentation algorithms. The cluster selection support unit 151 outputs an image indicating the candidate clusters by applying visual effects such as color, pattern, or blinking to clusters with small distance scales. For example, when one cluster is selected by the user, the cluster selection support unit 151 may indicate clusters with small distance scales as candidate clusters based on the selected cluster, or it may indicate multiple clusters with small distance scales as candidate clusters when no cluster is selected.

[0125] An example of processing by the information processing device according to Embodiment 4 of this disclosure will be described. First, let's explain an example of a process for changing the number of segments. Figure 30 is a flowchart showing an example of a process for changing the number of segments in an information processing device according to Embodiment 4 of this disclosure. The process shown in Figure 30 is a process in which the process according to this embodiment is added to the process shown in Figure 24. The process shown in Figure 30 is a process in which a segment selection support process (step ST4130) is inserted between the process of step ST3120 shown in Figure 24 (the process of step ST4120 shown in Figure 30) and the process of step ST3130 (the process of step ST4140 shown in Figure 30). The information processing device 100D determines that it has received a segment number change command in the segment number change command reception determination process in step ST4110 (step ST4110 "Segment number change command received?" "YES"), executes the time series division result acquisition process in step ST4120, and then executes the segment selection support process (step ST4130). In the segment selection support process, the segment selection support unit 152 of the selection support unit 150 of the information processing device 100D uses the time series division results to determine the distance scale of features between segments derived from clustering and segmentation algorithms. The segment selection support unit 152 extracts candidate segments based on the distance scale of features between segments. The segment selection support unit 152 outputs an image indicating the extracted candidate segments. In contrast, by the user operating the input source device 10 to select and specify a segment, the information processing device 100D can execute the processing from step ST4130 onwards in the same manner as the processing from step ST3130 onwards shown in Figure 24.

[0126] Next, we will explain an example of a process for changing the number of clusters. Figure 31 is a flowchart showing an example of a process for changing the number of clusters in an information processing device according to Embodiment 4 of this disclosure. The process shown in Figure 31 is a process in which the process according to this embodiment is added to the process shown in Figure 25. The process shown in Figure 31 is a process in which a cluster selection support process (step ST4230) is inserted between the process of step ST3220 shown in Figure 25 (the process of step ST4220 shown in Figure 31) and the process of step ST3230 (the process of step ST4240 shown in Figure 31). The information processing device 100D determines that it has received a cluster number change command in the cluster number change command reception determination process in step ST4210 (step ST4210 "Cluster number change command received?" "YES"), executes the clustering result acquisition process in step ST4220, and then executes the cluster selection support process (step ST4230). In the cluster selection support process, the cluster selection support unit 151 of the selection support unit 150 of the information processing device 100D uses the clustering results to determine the distance scale of features between segments derived from the clustering and segmentation algorithms. The cluster selection support unit 151 extracts candidate clusters based on the distance scale of features between clusters. The cluster selection support unit 151 outputs an image indicating the extracted candidate clusters. In contrast, by the user operating the input source device 10 to select and specify a cluster, the information processing device 100D can execute the processing from step ST4230 onwards in the same manner as the processing from step ST3230 onwards shown in Figure 25.

[0127] According to the configuration described above, it is possible to assist the user in suggesting the most suitable selection options when making a selection. Furthermore, it can visually indicate the optimal candidate segments and clusters, making them easier for users to understand.

[0128] This embodiment further illustrates an example of a configuration that can be expressed as follows. [4] A cluster selection support unit extracts candidate clusters based on the characteristics of each cluster in the clustering results (feature distance, distance based on features, distance in feature space) and outputs an image indicating the candidate cluster. The information processing apparatus according to [3], characterized by comprising [3]. This disclosure further provides an information processing device that can assist in selecting the optimal cluster when changing the number of clusters. Furthermore, by applying the above-mentioned information processing device to a time-series segmentation device, it is possible to provide a time-series segmentation device that, similarly to the above, enables the maintenance of clustering consistency when modifying clustering results. Furthermore, by applying the above configuration to a system including the above information processing device, the above information processing method, or the above program, etc., this disclosure will produce the same effects as described above.

[0129] This embodiment further illustrates an example of a configuration that can be expressed as follows. [8] A segment selection support unit extracts candidate segments based on the characteristics of each segment (feature distance, distance based on features, distance on features) in the aforementioned time-series segmentation results, and outputs an image indicating the candidate segment. The information processing apparatus according to [7], characterized by comprising [7]. This disclosure further provides an information processing device that can assist in selecting the optimal segment when changing the number of segments. Furthermore, by applying the above-mentioned information processing device to a time-series segmentation device, it is possible to provide a time-series segmentation device that, similarly to the above, enables the maintenance of clustering consistency when modifying clustering results. Furthermore, by applying the above configuration to a system including the above information processing device, the above information processing method, or the above program, etc., this disclosure will produce the same effects as described above.

[0130] Here, we will describe the hardware configuration required to realize the functions of this disclosure. Figure 32 shows a first example of a hardware configuration for realizing the functions of the configuration of this disclosure. Figure 33 shows a second example of a hardware configuration for realizing the functions of the configuration of this disclosure. Each of the information processing devices (time-series segmentation devices) 100 (100', 100A, 100B, 100C, 100D) disclosed herein is implemented by hardware as shown in Figure 32 or Figure 33.

[0131] The information processing device 100 (100', 100A, 100B, 100C, 100D) is composed of, for example, a processor 10001, a memory 10002, an input / output interface 10003, and a communication circuit 10004, as shown in Figure 32. Processor 10001 and memory 10002 are, for example, components installed in a computer. The memory 10002 contains the computer as follows: time-series division unit 101, clustering unit 102, change command reception unit 110, cluster number change command reception unit 111, segment number change command reception unit 112, cluster number change unit 120, cluster merging unit 121, cluster division unit 122, segment number change unit 130, segment merging unit 131, segment division unit 132, selection reception unit 140, selected cluster reception unit 141, selected segment reception unit 142, selection support unit 150, cluster selection support unit 151, segment selection support unit 152, Furthermore, a program for functioning as a control unit (not shown) is stored. The processor 10001 reads and executes the program stored in memory 10002, thereby enabling the time series division unit 101, clustering unit 102, change command reception unit 110, cluster number change command reception unit 111, segment number change command reception unit 112, cluster number change unit 120, cluster merging unit 121, cluster division unit 122, segment number change unit 130, segment merging unit 131, segment division unit 132, selection reception unit 140, selected cluster reception unit 141, selected segment reception unit 142, selection support unit 150, cluster selection support unit 151, segment selection support unit 152, Furthermore, the functions of a control unit (not shown) are realized. Furthermore, a storage unit (not shown) is realized by memory 10002 or other memory (not shown). Furthermore, a communication unit (not shown) is realized by the communication circuit 10004.

[0132] Processor 10001 uses, for example, a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a microprocessor, a microcontroller, or a DSP (Digital Signal Processor). Memory 10002 may be a non-volatile or volatile semiconductor memory such as RAM (Random Access Memory), ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable Read Only Memory), or flash memory; it may be a magnetic disk such as a hard disk or flexible disk; it may be an optical disk such as a CD (Compact Disc) or DVD (Digital Versatile Disc); or it may be a magneto-optical disk. The processor 10001 and the memory 10002 or the communication circuit 10004 are connected in a way that allows them to transmit data to each other. Furthermore, the processor 10001, the memory 10002, and the communication circuit 10004 are connected to other hardware via the input / output interface 10003 in a way that allows them to transmit data to each other.

[0133] Alternatively, the functions of the time-series division unit 101, clustering unit 102, change command reception unit 110, cluster number change command reception unit 111, segment number change command reception unit 112, cluster number change unit 120, cluster merging unit 121, cluster division unit 122, segment number change unit 130, segment merging unit 131, segment division unit 132, selection reception unit 140, selected cluster reception unit 141, selected segment reception unit 142, selection support unit 150, cluster selection support unit 151, segment selection support unit 152, and control unit (not shown) in the information processing device 100 (100', 100A, 100B, 100C, 100D) may be realized by a dedicated processing circuit 20001, as shown in Figure 33.

[0134] The processing circuit 20001 may utilize, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), a PLD (Programmable Logic Device), an FPGA (Field-Programmable Gate Array), a SoC (System-on-a-Chip), or a system LSI (Large-Scale Integration). Furthermore, a storage unit (not shown) is realized by memory 20002 or other memory (not shown). Memory 20002 may be a non-volatile or volatile semiconductor memory such as RAM (Random Access Memory), ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable Read Only Memory), or flash memory; it may be a magnetic disk such as a hard disk or flexible disk; it may be an optical disk such as a CD (Compact Disc) or DVD (Digital Versatile Disc); or it may be a magneto-optical disk. Furthermore, a communication unit (not shown) is realized by the communication circuit 20004. The processing circuit 20001 and the memory 20002 or the communication circuit 20004 are connected in a way that allows them to transmit data to each other. Furthermore, the processing circuit 20001, the memory 20002, and the communication circuit 20004 are connected in a way that allows them to transmit data to other hardware via the input / output interface 20003. Note that the information processing device 100 (100', 100A, 100B, 100C, 100D) The functions of the time-series division unit 101, clustering unit 102, change command reception unit 110, cluster number change command reception unit 111, segment number change command reception unit 112, cluster number change unit 120, cluster merging unit 121, cluster division unit 122, segment number change unit 130, segment merging unit 131, segment division unit 132, selection reception unit 140, selected cluster reception unit 141, selected segment reception unit 142, selection support unit 150, cluster selection support unit 151, segment selection support unit 152, and the control unit (not shown) may be implemented by separate processing circuits, or they may be implemented together in a single processing circuit.

[0135] Alternatively, information processing device 100 (100', 100A, 100B, 100C, 100D) In this configuration, some functions of the time-series division unit 101, clustering unit 102, change command reception unit 110, cluster number change command reception unit 111, segment number change command reception unit 112, cluster number change unit 120, cluster merging unit 121, cluster division unit 122, segment number change unit 130, segment merging unit 131, segment division unit 132, selection reception unit 140, selected cluster reception unit 141, selected segment reception unit 142, selection support unit 150, cluster selection support unit 151, segment selection support unit 152, and control units (not shown) may be implemented by the processor 10001 and memory 10002, while the remaining functions are implemented by the processing circuit 20001.

[0136] Within the scope of this disclosure, it is possible to freely combine the embodiments, modify any component of each embodiment, or omit any component of each embodiment. [Industrial applicability]

[0137] This disclosure makes it possible to maintain the consistency of clustering when changing the number of clusters in the clustering results, and is therefore suitable for use in information processing devices or information processing systems that have a time series segmentation function, for example. By applying the information processing device or information processing system of this disclosure to time series data acquired from various devices, for example, the consistency of the clustering of the time series data is maintained, which can be used to improve the accuracy of anomaly detection in devices and improve the operating efficiency of devices. [Explanation of symbols]

[0138] 1(1A,1B,1C,1D) Information processing system (time series segmentation system), 10 Input source device (input / output device), 20 Output destination device (input / output device), 30 Input / output device, 100(100',100A,100B,100C,100D) Information processing device (time series segmentation device), 101 Time series division unit, 102 Clustering unit, 110 Change command reception unit, 111 Cluster number change command reception unit, 112 Segment number change command reception unit, 120 Cluster number change unit, 121 Cluster merging unit, 122 Cluster division unit, 130 Segment number change unit, 131 Segment merging unit, 132 Segment division unit, 140 Selection reception unit, 141 Selected cluster reception unit, 142 Selected segment reception unit, 150 Selection support unit, 151 Cluster selection support unit, 152 Segment selection support unit, 1100 Time series data, 1200 Segment (segment group), 1300 Cluster (cluster group), 1500 Output image (screen), 1510 Time series data, 1520 Clustering result, 1521 Cluster, 1522 Cluster, 1530 Cluster count command operation unit, 1531 Command cluster count display unit, 1540 Clustering target data selection unit, 1600 Output image (screen), 1610 Time series data, 1620 Clustering result, 1621 Cluster, 1630 Cluster count command operation unit, 1631 Command cluster count display unit, 1640 Clustering target data selection unit, 10001 Processor, 10002 Memory, 10003 Input / Output interface, 10004 Communication circuit, 20001 Processing circuit, 20002 Memory, 20003 Input / output interface, 20004 communication circuit.

Claims

1. An information processing device that modifies the number of clusters in a clustering result, which is the result of dividing time-series data into segments and clustering it, A change command receiving unit that receives a command to change the number of clusters, which indicates the commanded number of clusters, A cluster number change unit, upon receiving the aforementioned cluster number change command, determines whether to merge or split clusters based on the clustering results and the cluster number change command, determines which clusters to merge or split based on the characteristics of each cluster in the clustering results, and outputs the clustering results after the cluster number has been changed. An information processing device characterized by having the following features.

2. The aforementioned cluster number changing unit is: A cluster partitioning unit divides the clusters to be modified from the aforementioned clustering results and maintains the clusters other than those to be modified. A cluster merging unit merges the clusters to be changed from the aforementioned clustering results, while maintaining the clusters that are not to be changed. Equipped with, The information processing apparatus according to feature 1.

3. A selected cluster receiving unit receives a selected cluster, which is a cluster selected from among the clusters in the clustering results. Equipped with, The cluster number changing unit performs cluster merging or splitting using the selected clusters. The information processing apparatus according to claim 1 or 2.

4. A cluster selection support unit extracts candidate clusters based on the characteristics of each cluster in the clustering results and indicates which cluster is to be selected. The information processing apparatus according to claim 3, characterized by comprising:

5. The aforementioned change command receiving unit further receives a segment number change command indicating the commanded number of segments, When the segment number change command is received, the segment number change unit determines and executes the merging or splitting of the segments, which are the result of the time series data being divided into segments, based on the time series division result and the segment number change command, and outputs the time series division result after the segment number has been changed. The information processing apparatus according to claim 1 or 2, further comprising:

6. The segment number changing unit is, A segment division unit divides the segments to be changed from the aforementioned time-series division results and maintains the segments other than those to be changed. A segment merging unit merges the segments to be changed from the clustering results and maintains the segments that are not to be changed, Equipped with, The information processing apparatus according to feature 5.

7. A selection segment receiving unit that receives a selection segment, which is a segment selected from the segments into which the time-series data has been divided. Equipped with, The segment number changing unit performs segment merging or splitting using the selected segments. The information processing apparatus according to feature 5.

8. A segment selection support unit extracts candidate segments based on the characteristics of each segment in the time-series segmentation results and instructs the selection of candidate segments. The information processing apparatus according to claim 7, characterized by comprising:

9. A time-series segmentation unit that divides the aforementioned time-series data into segments, A clustering unit that outputs the clustering result, which is the result of clustering based on the respective characteristics of each segment, Equipped with, The cluster number change unit determines and executes the merging or splitting of the clusters, which are the clustering results, based on the clustering results output by the clustering unit and the cluster number change command, and outputs the clustering results after the cluster number has been changed. The information processing apparatus according to claim 1 or 2.

10. An information processing method by an information processing device that changes the number of clusters in a clustering result, which is the result of clustering time series data, The change command receiving unit of the information processing device receives a cluster number change command indicating the commanded number of clusters, When the cluster number change command is received, the cluster number change unit of the information processing device decides whether to merge or split the clusters based on the clustering results and the cluster number change command, determines which clusters to merge or split based on the characteristics of each cluster in the clustering results, and outputs the clustering results after the cluster number has been changed. An information processing method characterized by the following:

11. Computers, An information processing device that modifies the number of clusters in a clustering result, which is the result of clustering time-series data, A change command receiving unit that receives a command to change the number of clusters, which indicates the commanded number of clusters, A cluster number change unit, upon receiving the aforementioned cluster number change command, determines whether to merge or split clusters based on the clustering results and the cluster number change command, determines which clusters to merge or split based on the characteristics of each cluster in the clustering results, and outputs the clustering results after the cluster number has been changed. A program characterized by operating as an information processing device equipped with [a specific feature].

Images