New Measure of Interestingness for Efficient Extraction of Association Rules

Data Mining helps to uncover the already unknown and non-redundant knowledge in large databases, which can be used for decision making purpose. Association rule mining is one of the key research area in the field of Data Mining. Association rule mining can be considered as unsupervised learning model, it discovers the interesting relationship among large set of data items on the basis of some predefined threshold. Support-confidence is the classical model used for the rule mining purpose, it uses confidence for final rule generation but it has some limitations. As sometimes it can generate those rules which are not positively correlated and thus can mislead the decision maker. In this paper we addressed the problems associated with existing approach and also proposed two new measure of interestingness to deal with these problems. The new measures have been tested for their correctness.