User Interactive PostProcessing of Association Rules and Correlation based Redundancy Removal
C. Sweetlin, V. Kalaivani · 2012
Traditional association rule mining generates a large number of rules. This leads to a difficulty in finding the interested and significant rules. An efficient interactive post-processing task which includes ontology and rule schema is used to obtain user interesting rules. Correlation analysis finds significant association rules by analyzing the dependency between the antecedent and consequent parts of the rule. In this paper, correlation analysis is integrated with the interactive postprocessing to obtain significant user interesting rules. A redundancy removal follows this framework to weed out the extra rules and also to reduce the ruleset further. The proposed methodology provides a significant set of non-redundant user interesting rules leading to an efficient analysis.