抽样技术在序列模式增量更新中的应用

在这篇文章中，我们提出了一种应用抽样的技术于序列挖掘的算法。这个方法能在原
始数据库和更新后的数据库之间评价序列模式的变化。评价了序列模式的变化我们能决定何时使用精确的挖掘算法或增量算法：如果变化达到一定程度，就使用精确的方法以挖掘新的序列模式；如果变化比较小，则在一定的可信度情况下，就使用原来的模式，且误差能控制在一定的范围内。
关键字：抽样，数据挖掘，序列，置信区间
Abstract : In this paper, we use sample method into sequence pattern mining and devise a
algorithm. This method can estimate the changes of patterns in a database when it is updated.
When we estimate the changes of patterns, we can make a decision about the time to use an exact mining algorithm or an incremental algorithm: if the changes reach a certain degree, exact method will be used in order to mine new sequence patterns; if the change is little, we can use the original patterns under a certain reliability, and the error can be controlled in certain bound.
Key words: sample, data mining, sequence, confidence interval.