在有噪音的环境中挖掘序列模式精简基

提出在匹配度模型下频繁序列模式精简基的概念,精简基由相对于一系列匹配度阈值的最大序列模式组成,它是频繁序列模式的一个子集,可以用它来估计任一频繁序列模式的匹配度,并能将误差控制在确定范围内.还开发了一个从有噪音的序列数据库中挖掘这种序列模式精简基的算法,该算法采用了一种不需要保留候选序列模式的方法来检查最大序列模式,采用的剪枝技术也比以前的算法更有效率.实验结果显示计算频繁序列模式精简基是很有前途的.

Study of mining condensed sequential pattern bases in a noisy environment

The concept of condensed frequent sequential pattern base under the match metric was proposed. The base was consisted of maximal frequent sequential patterns for a series of match thresholds. It was a subset of frequent sequential patterns, and was used to estimate the match of arbitrary frequent sequential patterns in a guaranteed maximal error bound. An algorithm was developed to mine such a condensed sequential pattern base from a noisy sequence database. The algorithm adopted a scheme to check maximal sequence patterns without candidate maintenance, and pruned the search space more deeply in comparison with the pervious algorithms. Experimental results showed that computation of condensed frequent sequential pattern base is promising.