一种时间序列案例的综合相似度函数

One Integrated Similarity Function for Temporal Serial Cases

针对案例推理中案例检索的相似度函数这个关键问题，本文分析了时间序列案例的结构特点，提出了一种时间序列案例的相似度函数。讨论了常用相似度函数的不足，提出了一种综合相似度函数。最后，提出了一种时间序列案例的综合相似度函数，并在气象数据上进行了对比检验与分析。结果表明。该函数不仅可行，而且比常用函数更优。

Similarity function is critical to similarity retrieval in case-based reasoning （CBR）.Based on one analysis of the structnral characteristics of temporal serial cases,this paper presented a similarity function for temporal serial cases.After discussed the shurtcomings of common similarity functions,an integrated similarity function was presented.Then a integrated similarity function for temporal serial cases was presented and was tested on meteorologic data.The results suggest that the function is not only feasible,but also performs much better than common functions.