Parallel mining association rules with bit string array in large database

Association rule mining is one of the important data mining tasks. However, the previously proposed methods still encounter some problems, such as complex data structure, candidate set generation, and so on. To improve efficiency, association rules can be mined in parallel. In this paper, we use a simpler data structure called bit string array and propose a new approach to apply parallel projection and compress technique in parallel mining association rules. It conducts various operations on bit string array according to the frequency of frequent items. For frequent item with less frequency, we conduct set operation on them; for frequent item with more frequency, we adopt compress technique and conduct bit AND operator on them. Moreover, it will reduce the communication cost and also response time. This method can be scaled up to very large databases by parallel projection and compress technique.