High-Efficiency Algorithm for Mining Maximal Frequent Item Sets Based on Matrix

Abstract

Association Rule Mining is an important data mining technique and Maximal frequent item sets mining is an essential step in the process of Association rule. Here presented is BM-MFI, a new algorithm based on matrix, for mining maximal frequent item sets. Its basic idea is transforming the event database into matrix database by operating the rows and columns of matrix to compress the database. Using Itemset-Tidset pair can mine maximal frequent item sets in the compressed database with convenience and effectiveness, and therefore prevent conditional FP-tree and candidate patterns. Experimental result verifies the efficiency of the BM-MFI.