Association Rule Pruning based on Interestingness Measures with Clustering

TL;DR

This paper explores how association rule clusters are distributed across various interestingness measures, aiming to improve rule pruning and grouping in knowledge mining.

Contribution

It introduces an analysis of rule clusters based on interestingness measures, enhancing understanding of rule distribution for better pruning strategies.

Findings

Rule clusters vary significantly across different interestingness measures.

Clustering helps identify more meaningful rules for knowledge discovery.

The study provides insights into effective rule pruning techniques.

Abstract

Association rule mining plays vital part in knowledge mining. The difficult task is discovering knowledge or useful rules from the large number of rules generated for reduced support. For pruning or grouping rules, several techniques are used such as rule structure cover methods, informative cover methods, rule clustering, etc. Another way of selecting association rules is based on interestingness measures such as support, confidence, correlation, and so on. In this paper, we study how rule clusters of the pattern Xi - Y are distributed over different interestingness measures.