| 基于DHP的频繁遍历路径挖掘算法 |
| |
| 引用本文： | 王涛伟,周必水. 基于DHP的频繁遍历路径挖掘算法[J]. 杭州电子科技大学学报, 2005, 25(5): 60-63 |
| |
| 作者姓名： | 王涛伟 周必水 |
| |
| 作者单位： | 1. 浙江万里学院,计算机系,浙江,宁波,315100
2. 杭州电子科技大学,计算机学院,浙江,杭州,310018 |
| |
| 摘 要： | 大量的候选项集是挖掘路径遍历模式中的主要问题.针对这个问题,提出了基于DHP算法的路径遍历算法,通过使用哈希技术在产生侯选项集时删除不满足条件的项目,这种方法特别是在产生候选2-项目时效率非常高,这样很好的解决了整个处理过程的性能瓶颈.另外,使用了整枝技术使事务数据库的大小在每次扫描后迅速减小.实验结果表明,基于DHP的频繁遍历路径算法在挖掘频繁项目集时是有效的.
|
| 关 键 词： | 频繁遍历路径 数据挖掘 序列模式 关联规则 候选项目集 整枝技术 |
| 文章编号： | 1001-9146（2005）05-0060-04 |
| 收稿时间： | 2005-03-04 |
| 修稿时间： | 2005-03-04 |
An Algorithm for Mining Frequent Path Traversal Based on DHP |
| |
| WANG Tao-wei,ZHOU Bi-shui. An Algorithm for Mining Frequent Path Traversal Based on DHP[J]. Journal of Hangzhou Dianzi University, 2005, 25(5): 60-63 |
| |
| Authors: | WANG Tao-wei ZHOU Bi-shui |
| |
| Affiliation: | 1. Department of Computer Science, Zhejiang Wardi University, Ningbo Zhejiang 315100 China ; 2. School of Computer Science, Hangzhou Dianzi University, Hangzhou Zhejiang 310018, China |
| |
| Abstract: | A huge set of candidates is the main problem in mining the path traversal patterns. In relation to this, the paper puts forward a path traversal algorithrn based on DHP, and uses a hashing technique to remove unsatisfied reference in the candidate itemsets generation. This method is effective especially in the number of candidate 2 - itemsets generation, hence resolving the performance bottleneck. In addition, pruning technique enables us to rapidly prone the transaction database size after every scan. Experimental results show that the algorithm is effective in mining frequent itemsets. |
| |
| Keywords: | frequent path traversal data mining sequential pattern association roles candidate itemsets pruning technique |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
