Performance testing based on variable length segmentation and clustering of time series data

• US 10,643,138 B2
• Filed: 01/30/2015
• Issued: 05/05/2020
• Est. Priority Date: 01/30/2015
• Status: Active Grant

• Current Assignee: Micro Focus LLC (Open Text Corporation)
• Original Assignee: Micro Focus LLC (Open Text Corporation)
• Inventors: Bellala, Gowtham, Zhang, Mi, Lyon, Geoff M.
• Primary Examiner(s): Smith, Paulinho E
• Application Number:
  US15/325,786
• Publication Number:
  US 20170147930A1
• Time in Patent Office: 1,922 Days
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/04 Inference or reasoning models

Abstract

Performance testing based on variable length segmentation and clustering of time series data is disclosed. One example is a system including a training module, a performance testing module, and an interface module. The training module generates a trained model to learn characteristics of a system of interest from training time series data by segmenting the training time series data into homogeneous windows of variable length, clustering the segments to identify patterns, and associating each cluster with a cluster score. The performance testing module analyzes system characteristics from testing time series data by receiving the testing time series data, and determining a performance metric for the testing time series data by analyzing the testing time series data based on the trained model. The interface module is communicatively linked to the performance testing module, and provides the performance metric via an interactive graphical user interface.

Claims

1. A device comprising:

a processor; and

a memory storing instructions that are executable to cause the processor to;

receive training time series data from a system of interest,

segment the training time series data into segments of homogeneous windows of variable lengths, wherein each segment of the segments of homogeneous windows of variable lengths is a segment of the training time series data that contains values that are comparable with each other within a predetermined threshold of deviation,

cluster the segments into clusters,

associate each cluster of the clusters with a cluster score,

generate a trained model including the clusters and the cluster scores,

receive testing time series data from the system of interest,

determine a performance metric for the testing time series data by analyzing the testing time series data based on the trained model, and

generate, for display, the performance metric on an interactive graphical user interface.

2. The device of claim 1, wherein to cluster the segments into clusters, the instructions cause the processor to categorize the segments into a plurality of bins based on the variable lengths of the segments.

3. The device of claim 1, wherein the cluster score of each cluster is generated based on predetermined rules.

4. The device of claim 1, wherein the instructions cause the processor to pre-process the testing time series data based on one of data interpolation, data imputation, feature sub-selection, data fusion, validation, estimation, and editing.

5. The device of claim 1, wherein the instructions cause the processor to:

generate additional segments of the testing time series data;

associate each of the additional segments with one of the clusters in the trained model; and

associate each of the additional segments with a segment score based on the cluster score for the associated cluster.

6. The device of claim 5, wherein the performance metric is based on a weighted aggregate of the segment scores.

7. The device of claim 1, wherein characteristics of the system of interest include one of operational and behavioral performance of the system.

8. The device of claim 7, wherein the performance metric is a behavior scorecard for a vehicle driver.

9. The device of claim 8, wherein the instructions cause the processor to associate each cluster of the clusters with a cluster score based on a variance of a load value.

10. A method comprising:

receiving training time series data of a system;

segmenting, by a processor, the training time series data into segments of homogeneous windows of variable lengths, wherein each of the segments of homogeneous windows of variable lengths is a segment of the training time series data that contains values that are comparable with each other within a predetermined threshold of deviation;

clustering, by the processor, the segments into training clusters;

associating, by the processor, each of the training clusters with a cluster score;

generating, by the processor, a trained model including the training clusters and the cluster scores;

receiving testing time series data from the system;

determining, by the processor, a performance metric for the testing time series data based on the trained model; and

providing the performance metric via an interactive graphical user interface.

11. The method of claim 10, wherein determining the performance metric for the testing time series data based on the trained model comprises:

segmenting the testing time series data into segments of homogeneous windows;

associating each of the segments of the testing time series data with one of the training clusters in the trained model; and

determining a segment score for each of the segments of the testing time series data based on one of the cluster scores in the trained model.

12. The method of claim 11, wherein determining the performance metric for the testing time series data is based on a weighted aggregate of the segment scores for the segments of the testing time series data.

13. The method of claim 12, wherein the segments of the testing time series data have variable lengths, and wherein the segment scores for the segments of the testing time series data are weighted based on the variable lengths of the segments.

14. The method of claim 10, further comprising:

pre-processing the testing time series data based on one of data interpolation, data imputation, feature sub-selection, data fusion, validation, estimation, and editing.

15. A non-transitory computer readable medium comprising executable instructions to cause a processor to:

receive training time series data from a plurality of sources;

segment the training time series data into segments of homogeneous windows of variable lengths, wherein each segment of the segments of homogeneous windows of variable lengths is a segment of the training time series data that contains values that are comparable with each other within a predetermined threshold of deviation;

cluster the segments into training clusters;

associate each of the training clusters with a cluster score;

generate a trained model including the training clusters and the cluster scores;

receive testing time series data from a system of interest;

determine a performance metric for the testing time series data by analyzing the testing time series data based on the trained model; and

provide the performance metric via an interactive graphical user interface.

16. The non-transitory computer readable medium of claim 15, wherein to determine the performance metric for the testing time series data based on the trained model, the instructions cause the processor to:

segment the testing time series data into segments of homogeneous windows;

associate each of the segments of the testing time series data with one of the training clusters in the trained model; and

determine a segment score for each of the segments of the testing time series data based on one of the cluster scores in the trained model.

17. The non-transitory computer readable medium of claim 16, wherein the instructions cause the processor to determine the performance metric for the testing time series data based on a weighted aggregate of the segment scores for the segments of the testing time series data.

18. The non-transitory computer readable medium of claim 17, wherein the segments of the testing time series data have variable lengths, and wherein the segment scores for the segments of the testing time series data are weighted based on the variable lengths of the segments.