ASSESSING THE MICROECONOMIC FACET OF ASSOCIATION RULES VIA AN EFFICIENT WEIGHTING SCHEME
Ioannis Kouris, Christos H. Makris, Athanasios K. Tsakalidis · 2004
AbstractMost algorithms and approaches dealing with the task of association rule mining have assumed all itemsets to be of the same nature and importance and used a single support. Very few have tried to address the non-uniformity and non-homogeneity of both the items and also their frequencies. Nevertheless none of the approaches that we are aware of have proposed a concrete way of identifying and assigning the correct measure of importance to every itemset; neither have they taken into consideration the framework within which a data mining activity should be viewed and implemented. In the paper we look into mining for retail organizations and view itemsets and associations rules through the appropriate microeconomic framework. We propose a weighting scheme that assigns the correct supports to all itemsets, automatically finds the most interesting ones and yet proves very efficient.