Sequential Data Classification via Confidence Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing information processing systems for biometric authentication, such as those using sequential probability ratio test (SPRT), face challenges in accurately classifying sequential data due to the lack of consideration for correlation between elements, leading to reduced classification accuracy, especially in time-series data with strong correlations.
Innovation Solution
An information processing system that includes an acquisition unit, a first calculation unit to determine initial indicators, a weight calculation unit to assign weights based on confidence, and a second calculation unit to calculate a weighted second indicator for classifying sequential data, allowing for enhanced classification by considering the correlation between elements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sequential probability ratio test (SPRT) is used for classification, then classification process can be implemented, but classification accuracy deteriorates due to lack of consideration for correlation between elements
Solution Approach 1:
The patent segments the classification process into multiple stages: element-level indicator calculation, weight calculation based on confidence, and sequential aggregation of indicators. This segmentation allows the system to handle correlation between elements by processing them individually first, then combining results with appropriate weights, thereby improving classification accuracy without overwhelming complexity.
Solution Approach 2:
The patent applies local quality by calculating confidence levels for each individual element and using these confidence levels to determine weights. This allows the system to differentiate between reliable and unreliable elements, giving higher importance to elements with higher confidence. By applying this locally at the element level, the system improves overall classification accuracy while maintaining manageable complexity through standardized confidence-weighted aggregation.
2Measurement precision
If all elements are treated equally in classification, then calculation process is simple, but classification accuracy deteriorates when elements have different confidence levels
Solution Approach 1:
The patent changes the parameter from uniform treatment of all elements to differentiated treatment based on confidence levels. By introducing confidence as a parameter and using it to calculate weights, the system transforms the classification approach from simple equal-weight averaging to confidence-weighted aggregation. This parameter change improves classification accuracy by accounting for element reliability while maintaining relatively simple calculation processes through standard weighting techniques.
Data Source
AI summary
An information processing system (10) includes: an acquisition unit (50) configured to sequentially acquire a plurality of elements included in sequential data; a first calculation unit (110) configured to calculate, for each of the plurality of elements, a first indicator indicating which one of a plurality of classes the element belongs to; a weight calculation unit (130) configured to calculate, for each of the plurality of elements, a weight according to a confidence related to calculation of the first indicator; a second calculation unit (120) configured to calculate, based on the first indicators each weighted with the weight, a second indicator indicating which one of the plurality of classes the sequential data belongs to; and a classification unit (60) configured to classify the sequential data as any one of the plurality of classes, based on the second indicator. According to such an information processing system, sequential data can be appropriately classified.