A New Weighted Association Rule Suitable for Causes and Effects Analysis

Abstract

A new weighted association rule algorithm is discussed. It is based on the mining algorithms of minwal(0) and minwal(w). The new algorithm can effectively mine the association rules which define some attributes as antecedent partial, while others as consequent partial. The new algorithm also can effectively mine the association rules with low support and high confidence, and these association rules have greater significant in some application.