An m-ary tree based Frequent Temporal Pattern (FTP) mining algorithm

Abstract

Frequent set mining (FSM), an important phase of association rule mining, is the process of generating frequent sets that satisfy a specified minimum support threshold. This paper explores FSM in temporal data domain or FTP mining and proposes an efficient algorithm for the same. Existing algorithms for FTP mining are based on a priori's level wise principle. In conventional or transactional data domain, a priori has been proven to suffer from the repeated scans limitation and has been succeeded by several algorithms that overcome the setback. The proposed algorithm eliminates a priori's repeated scans limitation in temporal domain, requiring only two overall scans of the original input. Experimental results demonstrate the significant improvements in execution time of the proposed algorithm as opposed to the a priori based one