期刊论文

中文

小型微型计算机系统

1000-1220

2016

37

6

1212-1216

61170017:国家自然科学基金 61370108:国家自然科学基金

本校为第一机构

随着高效用模式挖掘算法在实际应用中的重要性逐步显著,因而其得到了越来越多的关注和研究,但是已知的一些算法存在着多遍数据集扫描以及会产生大量候选项集、时效性不高等问题. 这些问题使得高效用模式的挖掘效率大大降低,故本文提出一种新的基于投影的高效用项集挖掘算法HUPMP(high utility pattern mining on projection). 采用基于投影的方式将项集的信息存储在数组中,能有效的减少前人基于树结构建树所产生的耗时. 同时HUPMP算法将多次数据集扫描减少为只需要进行两次数据集的扫描便可挖掘出所有的高效用模式. 另外,针对会产生大量候选项集的问题,提出了一种不同于闭包属性的新性质.通过与...

High utility pattern mining algorithm plays a more and more important role in practical application, which has been widely studied. However, there are many problems in the existing algorithms, such as multiple database scanning, a large number of candidate itemsets and time consuming. These problems make the efficiency of high utility pattern mining is greatly reduced,and so this paper presents a new high utility itemsets mining algorithm based on projection called HUPMP（ high utility pattern mining on projection}. U- sing projection based app...