An Efficient Load Balancing Multi-core Frequent Patterns Mining Algorithm

Mining frequent pattern from transactional database is an important problem in data mining. Many methods have been proposed to solve this problem. However, the computation time still increase significantly while the data size grows. Therefore, parallel computing is a good strategy to solve this problem. Researchers have proposed various parallel and distributed algorithms on cluster system, grid system. However, the construction and maintenance cost is pretty high. In this paper, a multi-core load balancing frequent pattern mining algorithm is presented. The main goal of the proposed algorithm is to reduce the massive duplicated candidates generated in previous method. In order to verify the performance, we also implemented the proposed algorithm as well as previous methods for comparison. The experimental results showed that our method could reduce the computation time dramatically with more threads. Moreover, we could observe that the workload was equally dispatched to each computing unit.