An efficient mining of dominant entity based association rules in multi-databases

Abstract

Today, we have a large collection of data that is organized in the form of a set of relations which is partitioned into several databases. There could be implicit associations among various parts of this data. In this paper, we give a scheme for retrieving these associations using the notion of dominant entity. We propose a scheme for mining for dominant entity based association rules (DEBARs) which is not constrained to look for co-occurrence of values in tuples. We show the importance of such a mining activity by taking a practical example called personalized mining. We introduce a novel structure called multi-database domain link network (MDLN) which can be used to generate DEBARs between the values of attributes belonging to different databases. We show that MDLN structure is compact and this property of MDLN structure permit it to be used for mining vary large size databases. Experimental results reveal the efficiency of the proposed scheme. � 2011 Springer-Verlag Berlin Heidelberg.