有无查询意图的分类与实现架构模型研究  

Research on Implementation Framework Model and Classification Based on Query Intention and Non-query Intention

作　　者：吴晓晖[1] 宋萍萍[1] 张荣欣[1] 

机构地区：[1]湖北汽车工业学院信息管理系,湖北十堰442002

出　　处：《情报科学》2009年第12期1829-1833,共5页Information Science

基　　金：湖北省教育厅基金(2008q238);湖北汽车工业学院基金(2008YQ23)

摘　　要：有无查询意图的句子和短语识别分类正确率的高低直接影响搜索引擎最后的检索结果的精度。针对目前中文搜索引擎中对查询句和短语缺乏理解﹑识别和分类的现状,同时结合前人对于中文问答系统的问句分类、句子分类、web搜索查询意图分类研究的基础上进行了进一步的拓展和集成研究,构建了基于有无查询意图的句子、短语的初步分类体系和设计了一款实现架构模型。The precise rate, which is higher or lower, on identification and classification of sentences and phrases based on query intention and non-query intention directly impacts the accuracy of final retrieval results of Search Engine. At present, the current situation of Chinese Search Engine is lack of understanding, identification, and classification for query sentences and phrases. Meanwhile, the further research is carried on based on the previous research, which is question classification on Chinese question-answering system and query-intention classification of web search. Finally, the preliminary taxonomy system based on query intention and non-query intention has been constructed and a implementation framework model is designed.

关 键 词：查询句理解 机器学习 问答系统 意图查询分类 支持向量机 

分 类 号：TP319.3[自动化与计算机技术—计算机软件与理论] G350[自动化与计算机技术—计算机科学与技术]