A Projection-Based Approach for Mining Highly Coherent Association Rules

In our previous approach, we proposed an apriori-based algorithm for mining highly coherent association rules, and it is time-consuming. In this paper, we present an efficient mining approach, which is a projection-based technique, to speed up the execution of finding highly coherent association rules. In particular, an indexing mechanism is designed to help find relevant transactions quickly from a set of data, and a pruning strategy is proposed as well to prune unpromising candidate itemsets early in mining. The experimental results show that the proposed algorithm outperforms the traditional mining approach for a real dataset.