Temporal Pattern Mining Using Discrepancy-Weighted Support
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data mining techniques for analyzing large medical datasets, such as electrocardiograms, face challenges in extracting relevant patterns due to the extraction of unnecessary patterns and increased computational complexity, particularly with higher maximum discrepancy values.
Innovation Solution
A method and apparatus for mining temporal patterns using an interest pattern model that generates data and candidate pattern groups based on a data suffix tree, calculates support values with varying weights, and determines pattern requirements to eliminate unnecessary patterns, optimizing mining speed and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If the maximum permissible discrepancy value is increased to extract more patterns, then the number of frequent patterns increases, but the computational complexity increases and unnecessary patterns are extracted
Solution Approach 1:
The patent segments the pattern extraction process into multiple stages: generating candidate patterns from data patterns, calculating support values with weights, and filtering based on minimum support thresholds. This segmentation allows the system to handle large datasets by processing patterns in manageable groups rather than computing all possible patterns simultaneously, thereby reducing computational complexity while maintaining the ability to extract a sufficient number of frequent patterns.
Solution Approach 2:
The patent applies preliminary filtering by generating candidate patterns with discrepancy values within permissible limits before performing full support value calculations. By pre-filtering candidate patterns based on discrepancy criteria and generating only promising candidates from data patterns, the system reduces the search space and computational burden while ensuring that the final set of frequent patterns meets the required quality standards.
2Quantity of substance
If the maximum permissible discrepancy value is increased to extract more patterns, then more patterns are extracted, but unnecessary patterns that do not actually appear in sequential data are extracted
Solution Approach 1:
The patent implements a feedback mechanism through the support value calculation that incorporates weights based on discrepancy values. Patterns are evaluated against actual sequential data occurrences, and the weighted support values provide feedback on pattern quality. This feedback loop ensures that patterns with high discrepancy values (which may be unnecessary) are penalized, while patterns that actually appear in the data are rewarded, thereby maintaining pattern accuracy even when extracting a large number of patterns.
Solution Approach 2:
The patent dynamically adjusts the evaluation parameters by applying weights to support values based on discrepancy levels. By changing the parameter weighting scheme - giving higher weights to patterns with lower discrepancy values and lower weights to patterns with higher discrepancy values - the system can control the balance between extracting more patterns and maintaining their accuracy. This parameter adjustment allows flexible control over pattern quality without requiring a fixed discrepancy threshold.
3Quantity of substance
If traditional data mining techniques are used to analyze large medical data, then data analysis can be performed, but the analytical capabilities of physicians are overwhelmed by the flood of medical data
Solution Approach 1:
The patent extracts only the most significant temporal patterns from large medical datasets by applying minimum support thresholds and discrepancy filters. Instead of presenting all possible patterns to physicians, the system extracts and prioritizes patterns that meet predetermined criteria for significance and reliability. This extraction process transforms the overwhelming flood of data into a manageable set of high-value insights that physicians can actually use for clinical decision-making.
Data Source
AI summary
An apparatus and a method for mining temporal pattern are provided. A method for mining temporal pattern includes generating a data pattern group comprising data patterns from sequential data, generating a candidate pattern group comprising candidate patterns from the data pattern group, calculating a support value for a candidate pattern in a candidate pattern group based on a discrepancy value of the candidate pattern, and determining whether the candidate pattern satisfies a predetermined pattern requirement, based on the calculated support value.