Phát triển các giải thuật song song trong khai phá luật kết hợp

In this paper we present two parallel algorithms for mining association rules that are well suited for distributed memory parallel computers. The algorithms are developed based on FP-growth method. The first algorithm is a task parallel formulation using a static load balancing technique. The second algorithm improves upon the first algorithm by dynamically balancing the load when the static task assignment leads to load imbalance.