Change-Point Detection Module Adjusting Data Values
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional change-point detection algorithms are inaccurate in identifying systematic changes in data due to real-world effects such as temporal, spatial, and seasonal fluctuations, which can lead to false detection of changes or data quality issues.
Innovation Solution
A change-point detection module that removes predetermined effects like temporal, spatial, and seasonal effects from data before applying change-point detection, using techniques such as decomposition into subsets, computation of mean values, and adjustment of data values to isolate systematic changes, and employing algorithms like quality control charts and cumulative sums for accurate detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional change-point detection algorithms are used to detect systematic changes in data, then the detection process is simple and straightforward, but the detection accuracy deteriorates due to false alarms caused by real-world effects such as temporal, spatial, and seasonal fluctuations
Solution Approach 1:
The patent segments the data into multiple subsets based on different time periods, spatial locations, or seasonal patterns. By dividing the data into distinct groups, the algorithm can analyze each subset separately to identify systematic changes while accounting for real-world effects like temporal and seasonal fluctuations, thereby improving detection accuracy without excessive complexity
Solution Approach 2:
The patent extracts and removes the effects of temporal, spatial, and seasonal variations from the data before performing change-point detection. By isolating and eliminating these confounding factors, the algorithm can focus on detecting true systematic changes, significantly reducing false alarms and improving measurement precision
2Reliability
If data is analyzed without removing predetermined effects, then the analysis process is fast and efficient, but the reliability of change detection deteriorates due to false detection of changes
Solution Approach 1:
The patent performs preliminary actions by pre-calculating and storing the effects of temporal, spatial, and seasonal patterns from historical data. These pre-computed effects are then subtracted from new data before change-point detection, allowing the system to maintain high reliability while reducing the time required for real-time analysis
Solution Approach 2:
The patent applies different adjustment strategies to different portions of the data based on local characteristics. By tailoring the effect removal process to specific time periods, locations, or patterns, the algorithm achieves high detection reliability while minimizing unnecessary processing steps and reducing overall computation time
Data Source
AI summary
To detect data change, data values are separated into plural sets. Predefined values representative of a predetermined effect are calculated for respective plural sets. Adjusted data values are calculated by removing impact of calculated predefined values from the data values in the respective plural sets. The data change is detected based on the adjusted data values.