Classification of association rules based on K-means algorithm

Abstract

Association rule mining is one of the most relevant techniques in data mining, aiming to extract correlation among sets of items or products in transactional databases. The huge number of association rules extracted represents the main obstacle that a decision maker faces. Hence, many interestingness measures have been proposed to evaluate the association rules. However, the abundance of these measures caused a new problem, which is the selection of measures that is best suited to the users. To bypass this problem, we propose an approach based on K-means algorithm to classify and to store Association Rules without favoring or excluding any measures. The experiments, performed on numerous datasets, show a significant performance of the proposed approach and it effectively classify the association rules.