An Efficient Algorithm for Association Mining

Abstract

Association rule discovery plays an important role in knowledge discovery and data mining, and efficiency is especially crucial for an algorithm to find frequent patterns from a large database. In this paper, a new algorithm called LogApriori algorithm is proposed by the idea of reducing unnecessary scanning of database in Apriori algorithm. The correctness of LogApriori algorithm is proved in this paper, and the performance of LogApriori algorithm is better than Apriori algorithm theoretically and practically. The success of LogApriori algorithm indicates that the strategy of producing itemsets with different number of items in one scanning can indeed find frequent patterns correctly and effectively.