CSPF: An Efficient Algorithm for Mining Consecutive Sequences

Abstract

Consecutively sequential patterns exist in many applications. The traversal paths in real world or in websites are kinds of consecutive sequences. Besides, the production flows of factories and the air or ocean currents are also kinds of consecutive sequences. However, few studies have been done on consecutively sequential patterns. In this paper we define the consecutive sequence and propose an algorithm CSPF (Mining Consecutively Sequential Patterns Using Filtering Approaches) to discover the consecutively sequential patterns. CSPF adopts two mechanisms to increase the performance of mining. First, CSPF algorithm uses a hash based technique, Hash MAP, which is similar to Hash Table to increase the access efficiency. Second, CSPF algorithm uses an infrequent consecutive sequences filtering mechanism to avoid generating a great deal of infrequent subsequences of transaction records. The CSPF algorithm only scans database four times and uses the filtration mechanism to filter out lots of non-frequent subsequences. The CSPF algorithm avoids wasting a lot of I/O time and increases the efficiency and the practicability in application. The experimental results indicate that the proposed CSPF algorithm outperform a significant improvement on mining consecutive sequences.