iFUM - Improved Fast Utility Mining

The main goals of Association Rule Mining (ARM) are to find all frequent itemsets and to build rules based of frequent itemsets. But a frequent itemset only reproduces the statistical correlation between items, and it does not reflect the semantic importance of the items. To overcome this limitation we go for a utility based itemset mining approach. Utility-based data mining is a broad topic that covers all aspects of economic utility in data mining. It takes in predictive and descriptive methods for data mining. High utility itemset mining is a research area of utility based descriptive data mining, aimed at finding itemsets that contribute most to the total utility. The well known faster and simpler algorithm for mining high utility itemsets from large transaction databases is Fast Utility Mining (FUM). In this proposed system we made a significant improvement in FUM algorithm to make the system faster than FUM. The algorithm is evaluated by applying it to IBM synthetic database. Experimental results show that the proposed algorithm is effective on the databases tested. General Terms Algorithms, Performance, Process, Results.