机构地区: 中国科学院计算技术研究所智能信息处理重点实验室
出 处: 《计算机科学》 2005年第4期31-33,共3页
摘 要: FP-growth算法是目前较高效的频繁模式挖掘算法之一。在FP-growth算法中,FP-树及条件FP-树的构造和遍历占了算法绝大部分的时间,如果能减少这方面的时间,则有望进一步改善算法的效率。本文给出了一个频繁模式挖掘算法SFP-growth。算法通过将FP-树有序化及采用高效排序算法等措施来提高FP-树构造的效率,从而使算法达到较高的效率。实验结果表明,SFP-growth是一个高效的频繁模式挖掘算法,其性能优于Apriori、Eclat和FP-growtn算法。 FP-growth is a high performance algorithm for mining frequent patterns. In FP-growth algorithm, it costs most of the time in constructing and traversing the FP-tree and conditional FP-tree. If we can reduce the time con- suming in tree construction and traversing, then the performance can be improved. In this paper, an improved algo- rithm, SFP-growth, is presented. The algorithm adopts sorted FP-trees to store the main information of the transac- tions. It also uses an efficient sorting algorithm and other techniques in the construction of trees. The experimental result shows that SFP-growth is an efficient algorithm, it outperforms Apriori, Eclat and FP-growth algorithm.