COBRA: Closed Sequential Pattern Mining Using Bi-phase Reduction Approach

Sequential pattern mining aims to find frequent patterns (guarded by a minimum support) in a database of sequences. As the support decreases the number of sequential patterns will increase rapidly. Therefore, a new trend is to mine closed sequential patterns, i.e. patterns which have no super-patterns with the same support in the database. Since mining closed sequential pattern has the same capability as mining the complete set of sequential patterns while reduces redundant patterns to be generated and stored, it is much economical and beneficial. In this paper we propose a novel approach which extends a frequent sequence with closed itemsets instead of single items. The motivation is that closed sequential patterns are composed of only closed itemsets. Hence, unnecessary item extensions which generates non-closed sequential patterns can be avoided. Furthermore, we propose LayerPruning which removes unnecessary enumeration by comparing candidate prefixes in the same layer rather than ...

• Type: Conference
• Year: 2006
• Where: DAWAK
• Authors: Kuo-Yu Huang, Chia-Hui Chang, Jiun-Hung Tung, Cheng-Tao Ho
• Keywords: Closed Sequential Patterns, DAWAK 2006, Information Management, Sequential Pattern Mining, Sequential Patterns