Cyclic Time-Series Anomaly Detection Consistency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The analysis of cyclic time-series data from finger-tapping motion systems for detecting brain dysfunction, such as dementia and Parkinson's disease, faces inconsistencies between overall data evaluation and partial data anomaly evaluation, leading to unreliable results due to discrepancies in anomaly rates and feature contribution levels.
Innovation Solution
A detecting apparatus and method that includes units for acquiring, processing, and analyzing cyclic information to detect anomalies in both overall and partial data, ensuring consistent evaluation by creating anomaly rates and feature importance levels, thereby providing reliable evaluations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If overall data evaluation is performed based on cyclic time-series data, then evaluation of entire data waveform is achieved, but inconsistency with partial data anomaly evaluation occurs
Solution Approach 1:
The patent segments the cyclic time-series data into multiple cycles and performs anomaly detection on each individual cycle. This segmentation allows the system to maintain both overall data evaluation and partial data anomaly evaluation with consistent results, resolving the contradiction between measurement precision and reliability.
2Difficulty of detecting and measuring
If anomaly detection is performed on partial data, then anomaly parts can be identified, but inconsistency with overall data evaluation results occurs
Solution Approach 1:
The patent implements a feedback mechanism where anomaly detection results from partial data are integrated back into the overall data evaluation. The system uses the anomaly detection results to adjust and refine the overall evaluation, ensuring consistency between partial and overall assessments while maintaining the ability to identify specific anomaly parts.
3Ease of operation
If different evaluation models are used for overall and partial data, then specific anomaly detection is enabled, but algorithm inconsistency arises
Solution Approach 1:
The patent employs a universal evaluation framework that can handle both overall data evaluation and partial data anomaly detection using the same underlying algorithm. This multi-functional approach eliminates algorithm inconsistency while maintaining the ability to provide clear anomaly identification, reducing system complexity despite the dual evaluation needs.
Data Source
AI summary
In the cyclic time-series data anomaly-part detecting system 1, the overall-data evaluating unit 13 detects the anomaly of the overall data 44A based on the overall-data feature amount 44B, and besides, creates the feature-amount importance level 45B. The anomaly-part detecting system 1 creates the partial data 46A by dividing the overall data 44A that is the cyclic time-series data, calculates the partial-data feature amount 46B, and displays and outputs the partial-data anomaly detection result 46C that is the detection result of the anomaly of the partial data 46A, based on the partial-data feature amount 46B and the feature-amount importance level 45B.