基于改进编辑距离的中文相似句子检索

中文相似句子检索的方法在基于实例的机器翻译等中文信息处理领域，具有非常广泛的应用背景。本文提出的基于改进编辑距离的中文相似句子检索方法，在使用信息检索技术提高检索效率的同时，以普通编辑距离算法为基础，加入了词汇的语义信息，使之更加符合中文句子相似度计算的要求。改进编辑距离与单纯基于语义辞典计算句子相似度的方法相比，具有便于扩展，准确率高等优点。在基于大规模双语句对检索的英文辅助写作系统中使用该算法进行中文句子检索，最后获得了81．33％的查准率和95．31％的查全率。

Similar Chinese Sentence Retrieval Based on Improved Edit-distance

The approach of similar Chinese sentence retrieval has been used widely in the field of Chinese information processing, such as Example Based Machine Translation (EBMT) and so on. This paper proposes the approach of similar Chinese sentence retrieval based on improved edit distance. It not only uses the technology of information retrieval to improve the efficiency of retrieval, but also adds the semantic information of words into the normal edit distance approach. The new approach is more consistent with the computation of Chinese sentence similarity. The approach of improved edit distance has more advantages than original edit distance algorithm, such as easily extending, high precision and so on. The new approach is used in the English writing assistant system based on a large bilingual sentences pairs and achieves 81 33% precision and 95 31% recall.