点击次数：

论文类型：期刊论文

发表时间：2015-01-01

发表刊物：ICIC Express Letters

收录刊物：EI、Scopus

卷号：9

期号：5

页面范围：1303-1310

ISSN号：1881803X

摘要：Mining association rules in medical data, is not only significant to the efficient use of existing medical data, but also significant to the discovery of new knowledge to improve health care standard. With the advent of big data era, it becomes more and more important to mine valuable association rules from massive data rapidly in order to meet real-time requirements. Traditional cluster computing based on MapReduce can achieve a good speed-up; however, the cost to construct a cluster is too high for ordinary users, which restricts its application range. In this paper, we propose a parallel Apriori algorithm based on multi-GPU (Graphics Processing Unit) to accelerate the calculation process of Apriori. The analytical results demonstrate that the proposed method can greatly speed up its execution with lower cost and provide ordinary users a feasible method in mining medical data. ? ICIC International