Mining closed sequential patterns using genetic algorithm

Closed sequential pattern mining is an important data mining task because it produces more compact result set and it is more efficient than sequential pattern mining. In general closed sequential patterns are generated from large data sets by applying algorithms like CloSpan and BIDE which require more execution time to compute all the closed sequential patterns. By using genetic algorithms we can reduce the execution time. The advantage of using a genetic algorithm in finding the closed sequential patterns is that it performs global search and it has less time complexity compared to other algorithms.This paper proposes a novel genetic algorithm G-CSPM to find closed sequential patterns. To improve the performance, we develop an effective fitness function and a pruning method. Our algorithm is the first method that utilizes genetic approach for closed sequential pattern mining. The results indicate that the proposed algorithm G-CSPM outperforms CloSpan.