A Parallel Mining Algorithm for Closed Sequential Patterns

Mining closed sequential patterns is an important data mining task with broad applications, the large dataset acquires us to use the parallel technique to solve the problems in data mining. A new parallel algorithm named Par-ClosP is introduced in this paper. It partitions the task to each processor, reduces the communication among the processors, uses pseudo projection technique to minimize the use of time and space, and it introduces a new pruning method, thus improves the efficiency of the algorithm.