Improved Frequent Pattern Mining Algorithm with Indexing

Abstract

Efficient frequent pattern mining algorithms are decisive for mining association rule. In this paper, we examine the matter of association rule mining for items in a massive database of sales transactions. Finding large patterns from database transactions has been suggested in many algorithms like Apriori, DHP, ECLAT, FP Growth etc. But here we have introduced newer algorithm called Improved Frequent Pattern Mining Algorithm with Indexing (IFPMAI), which is efficient for mining frequent patterns. IFPMAI uses subsume indexes i.e. those itemsets that co-occurrence with representative item can be identified quickly and directly using simple and quickest method. This will become beneficial like (i) avoid redundant operations of itemsets generation and (ii) many frequent items having the same supports as representative item, so the cost of support count is reduced hence the efficiency is improved. Then an example is used to illustrate this proposed algorithm. The results of the experiment show that the new algorithm in performance is more remarkable for mining frequent patterns.