High Qualified Pattern Mining Using Weight Occupancy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing high qualified pattern mining methods in transaction databases often result in irrelevance and incompleteness of recommended item sets due to reliance solely on support degrees, leading to poor user experience and misguidance in transactions.
Innovation Solution
A method and apparatus that determine high qualified patterns by considering both support degrees and weight occupancies, ensuring the support degree is greater than a preset threshold and the weight occupancy is above a minimum, to accurately identify relevant item sets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If only support degree is used to determine high qualified patterns, then the mining process is simple, but the accuracy of mining results deteriorates due to irrelevance and incompleteness of recommended item sets
Solution Approach 1:
The patent introduces a new parameter 'weight occupancy' in addition to the existing 'support degree' parameter. This changes the parameter set used for pattern evaluation, allowing the system to filter out low-quality patterns that have high support degree but low weight occupancy, thereby improving mining result accuracy without significantly complicating the overall process
Solution Approach 2:
The patent combines two different metrics (support degree and weight occupancy) into a composite evaluation framework. By requiring patterns to satisfy both conditions simultaneously, the system achieves more accurate and relevant pattern mining results while maintaining a relatively simple dual-threshold implementation approach
2Measurement precision
If multiple criteria (support degree and weight occupancy) are used to determine high qualified patterns, then the accuracy of mining results is improved, but the device complexity increases
Solution Approach 1:
The patent applies a partial action approach by introducing only one additional criterion (weight occupancy) rather than comprehensively redesigning the entire evaluation system. This dual-criteria approach provides sufficient improvement in accuracy while avoiding excessive complexity that would result from more comprehensive multi-dimensional evaluation frameworks
Data Source
AI summary
A high qualified pattern mining method is provided. The high qualified pattern mining method is performed by one or more processors, and includes determining a support degree of an item set to be processed in a transaction database; determining a weight occupancy of the item to be processed in the transaction database; and determining that the item set to be processed is a high qualified pattern based on the support degree not being less than a product of a preset lowest support degree threshold value and a total number of transactions in the transaction database, and the weight occupancy not being less than a preset minimum weight occupancy threshold value.