The improvement of weighted association rules arithmetic based on FP-tree

Abstract

When digging association rules among items, the items are dealt in an equal way. However, it is usually not happen in databases in the real world. Different items always have different importance. To reflect them, The way of draw weight into items and use weight association rules can solve the problem. But weighted association rules arithmetic of these research based on Apriori arithmetic at the present, There are problems as these: the candidate itemsets is huge, the cost of resources is high and the efficiency is low. In regard to these problems, a new Weight Association Rules method(FP-WAR) was discussed. FP-WAR based on FP-tree to discover weighted association rules, FP-WAR handle weight clipping technology based on k-support bound of item sets to improve markedly efficiency of arithmetic. The experiment shows that the efficiency of FP-WAR arithmetic is much better than current way to discover weighted association rules.