Time-Series Shapelet Classification With Existence Region Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing shapelet learning methods for time-series data struggle with specifying the region where shapelets appear, leading to inefficient classification performance and impractical calculation times, and lack techniques for visualizing the existence regions of shapelets.
Innovation Solution
An information processing device adjusts distances between time-series data and shapelets using a penalty degree to estimate existence regions, integrating multiple distance components and updating shapelets and classifiers to reduce classification loss, while visualizing these regions for improved classification accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If shapelet learning methods are used to classify time-series data, then classification performance can be improved, but it becomes difficult to specify the region where shapelets appear and calculation time becomes impractically long
Solution Approach 1:
The patent segments the time-series data by identifying specific regions where shapelets appear (existence regions). Instead of searching for shapelets across the entire time-series data, the method divides the search space into meaningful segments based on the detected existence regions, thereby reducing calculation time while maintaining classification performance.
Solution Approach 2:
The patent performs preliminary identification of shapelet existence regions before conducting the full shapelet learning process. By pre-specifying where shapelets are likely to appear in the time-series data, the method avoids unnecessary calculations in regions where shapelets do not exist, thus reducing overall calculation time.
2Measurement precision
If shapelet learning methods are used to classify time-series data, then classification performance can be improved, but the region where shapelets appear cannot be specified
Solution Approach 1:
The patent introduces an intermediary component that detects and identifies the existence regions of shapelets in time-series data. This intermediary step provides the crucial information about where shapelets appear, which is then used to guide the shapelet learning process and improve both interpretability and performance.
Solution Approach 2:
The patent replaces the traditional mechanical search approach with an information-based approach that uses detected existence region information to guide the shapelet learning process. Instead of brute-force searching through all possible regions, the method substitutes this with a guided approach based on identified existence regions.
3Device complexity
If traditional classification methods are used, then calculation is simpler, but classification performance and interpretability are insufficient
Solution Approach 1:
The patent segments the complex shapelet learning process into manageable steps: identifying existence regions, extracting shapelets from these regions, and performing classification. This segmentation makes the overall process more understandable and interpretable while maintaining high classification performance.
Data Source
AI summary
An information processing device includes one or more processors. The one or more processors are configured to: adjust a first distance between first time-series data and a plurality of identification patterns corresponding to a plurality of classes by using a degree of existence indicating a degree by which a plurality of times of the first time-series data are times at which the plurality of respective identification patterns exist; calculate a first feature amount indicating a feature of the first time-series data by using the adjusted first distance; and obtain, by inputting the first feature amount to an estimation model learning a parameter, the similar identification pattern, and the degree of existence to reduce an estimation error, the identification pattern similar to the first time-series data, the class, and the degree of existence.