Abstract: This paper studies the key problems of Uyghur, Kazak, Kyrgyz full-text search engine retrieval server and proposes an effective solution, including inputting Uyghur, Kazak, Kyrgyz keywords and shows normal search results without installing the local input method in users’ computers, detecting and correcting Uyghur, Kazak, Kyrgyz keywords with high spelling mistake rate, returning the correct and comprehensive search results to users, etc. Experimental results indicat that the solution provides convenience for users and obviously improves the precision and recall of Uyghur, Kazak, Kyrgyz search engine.

Key words: online processing, error detection, error correction, root segmentation, assimilation processing

摘要： 研究维、哈、柯全文搜索引擎检索器的关键问题，提出有效的解决方法，包括在用户计算机没有安装本地输入法和字库的情况下输入维、哈、柯文检索词并正常显示搜索结果，针对具有高拼写错误率的维、哈、柯文检索词进行检错、纠错处理，返回给用户正确而全面的搜索结果等。实验结果表明，该方法为用户提供方便的同时明显提高了维、哈、柯文搜索引擎的查全率和查准率。

关键词: 在线处理, 检错, 纠错, 词根切分, 同化处理

CLC Number: 

• TP311.13