A Novel Utility Sentient Approach for Mining Interesting Association Rules

Abstract

Utility-based data mining is a new research area that concentrates on all types of utility factors in data mining processes and is targeted at incorporating utility considerations in both predictive and descriptive data mining tasks. Discovering interesting association rules that are utilized to improve the business utility of an enterprise has long been recognized in data mining community. This necessitates identifying interesting association patterns that are both statistically and semantically important to the business utility. Classical association rule mining techniques are capable of identifying interesting association patterns but they have failed to associate the user's objective and utility in mining. In this paper, we have proposed an approach for mining novel interesting association patterns from transaction data items of an enterprise to improve its business utility. The approach mines novel interesting association patterns by providing importance to significance, utility and subjective interestingness of the users. The novel interesting patterns mined using proposed approach can be used to provide valuable suggestions to the enterprise to improve its business.