摘要： 提出了一种基于约束满足的启发式搜索方法，用于寻找时序数据相似变化区域。该算法利用所给出的一个局部约束条件和３个全局约束条件，从整体上模仿人对相似时序数据的模糊判断过程，以此找出数据序列中相似的变化区域。并给出了该算法对太阳黑子和纽约证券交易Ｃｏｍｐｏｓｉｔｅ指数数据的实验测试结果。

关键词: 相似时序数据, 约束满足, 启发式搜索, 数据挖掘

Abstract: Time-series similar search is an important study of data mining. This paper puts forward a new method based on the constlaint satisfactory search which combines human heuristic used in judging similarity in the time-series data. A local constraint and three global constraints are seven out to help search out the similar area existed in the bine-series with fuzzyness, and the experiments made on the practical time-series data are also given to prove its effectiveness on finding out similar time-series data.

中图分类号: 

• TP301.6