一种支持时间序列数据的CBR检索算法

探讨了如何为CBR（基于范例的推理）增加对一种特殊的范例类型——时间序列数据的支持．分析了基于谱分析的时间序列相似度比较算法不适用于CBR检索的缺点，并在此基础上设计了一种综合性能很好的CBR检索算法．思路是把时间序列相似度比较转化成一个卷积问题，并用DFT来简化这个卷积的计算．通过对这种CBR检索算法进行了深入的理论分析和认真的实验，结果证明，提出的算法是一个高效的算法．在这个检索算法的基础上，CBR就能够席用到时序数据的分析推理中，具有广阔的应用前景．

A CBR algorithm supporting time series data

This paper focuses on the retrieval algorithms of a special kind of CBR system in which cases are composed of time-series data. We introduced the classical algorithm used for processing similarity queries on time series data. This algorithm is based on the fact that DFT preserves the Euclidean distance in the time or frequency domain, and only the first few elements of the frequency sequence are significant, so the retrieval process can only use these significant elements to compute similarity degree. However, this algorithm has several disadvantages limiting its usage in CBR retrieval, so a new algorithm is presented for using batch method to compute the similarity degree. It is based on the observation that the original problem can be transformed to a convolution problem, and the circular convolution can be computed more efficiently using FFT. Theoretical analysis and experiment result prove that this algorithm is efficient and robust. The algorithm presented in this paper furnishes the CBR with the ability to process cases consist of time-series data.