摘要： 在基于K-means的案例检索算法中，目标案例初次映射的失败会导致案例检索成功率降低。针对该问题，提出一种基于滑动窗口的案例检索算法。分析滑动窗口维护策略，利用滑动窗口收集案例库中权重较高且最近常使用的案例，增加案例采样数量。实验结果表明，该算法能提高检索成功率，检索时间较短且案例映射次数较少。

关键词: 案例检索, 滑动窗口, 案例权重, 相似度

Abstract: The problem of low success rate of case retrieval is caused by the first mapping failure of target case in case retrieval algorithm based on the K-means clustering. Aiming at this problem, this paper proposes a case retrieval algorithm based on the sliding window and analyzes the maintain strategy of sliding window. It utilizes the sliding window to collect the cases which have high weight and recently the most frequency of use so that it adds the sampling number of case and improve the speed of case retrieval. Experimental results show that the algorithm can improve the success rate of retrieval, and obtain the better performance at the aspect of the retrieval time and the mapping number of target case.

Key words: case retrieval, sliding window, case weight, similarity

中图分类号: 

• TP393