Multi-Level Weighted Sequential Pattern Mining Based on Prime Encoding

Encoding can express the hierarchical relationship in the area of mining the multi-level sequential pattern, up to now all the algorithms of which find frequent sequences just according to frequency, but items have different importance in the real applications, therefore the weight constraint involved to the entire mining process is crucial. The MWSP algorithm based on the candidate generation-and-test approach is one of the best algorithms of the weighted sequential pattern mining, however, which will easily generate the situation of candidate combinatorial explosion during the mining process. Therefore, this paper presents the algorithm PMWSM, which adopts prime encoding to decide the parent-child relationship between different levels by one step of division operation, introduces the concept of K-minimum weighted support count to push weight constraint into the multi-level sequential pattern mining, utilizes the principle of prefix projection database to avoid the occurrence of candidate combinatorial explosion, and takes full advantage of the minimum weighted support count to optimize the algorithm. The experimental results show that the algorithm PMWSM is more effective than the algorithm MWSP on mining multi-level weighted sequential patterns from the sequence database.