Frequent Item-Set Mining Using Absence-Based Association Rules
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current computational resources are insufficient to effectively analyze large datasets for data mining, particularly when using exponential algorithms like frequent item-set mining, leading to inefficiencies in processing time and resource utilization due to the large volume of output data generated.
Innovation Solution
A system and method for frequent item-set mining that includes a support threshold and a reverse support threshold to filter items, generate negated items for absent items, and calculate absence-based association rules, reducing computational complexity and improving performance by focusing on meaningful correlations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If frequent item-set mining algorithms are used to analyze large datasets, then association rules can be discovered, but the computational resources and processing time become insufficient due to the exponential nature of the algorithms
Solution Approach 1:
The patent extracts and focuses only on the most promising itemsets by applying a reverse support threshold filter. Instead of generating all possible association rules, the system extracts only those itemsets where the absent item has high support (appears in most transactions), thereby reducing the search space from exponential to manageable size while preserving the most valuable association rules.
Solution Approach 2:
The patent inverts the traditional frequent item-set mining approach by focusing on absent items rather than present items. Instead of finding itemsets that frequently appear together, the system identifies itemsets where a specific item is frequently absent, generating negated items and association rules based on absence patterns. This inversion dramatically reduces computational complexity.
2Loss of information
If the volume of input data is increased to improve analysis comprehensiveness, then more patterns can be detected, but the output data volume becomes too large to process effectively
Solution Approach 1:
The patent changes the parameter used for filtering itemsets from traditional support threshold alone to a combination of support threshold and reverse support threshold. By requiring that absent items have support greater than a reverse threshold (e.g., 0.5), the system dramatically reduces the number of candidate itemsets generated, thereby reducing output data volume while maintaining comprehensive pattern detection for absence-based associations.
3Productivity
If traditional support threshold filtering is used, then itemsets can be filtered, but the resulting association rules may not be the most interesting or useful
Solution Approach 1:
The patent inverts the filtering criterion by focusing on absent items with high support rather than present items. By generating negated items for absent items and creating association rules based on absence patterns, the system discovers more interesting and useful rules such as 'absence of item A implies presence of item B', which provide different and often more actionable insights than traditional presence-based rules.
Data Source
AI summary
A dataset of a plurality of transactions is received, the plurality of transactions including at least a first item and a second item. An absence-based association rule correlating an absence of the first item with an inclusion of the second item within the plurality of transactions is calculated.