隐含语义索引在农业技术问答系统中的应用  

Application of Latent Semantic Index in Question-Answering System about Agriculture Technology

作　　者：魏保子[1] 王儒敬[1] 

机构地区：[1]中国科学院合肥智能机械研究所

出　　处：《微电子学与计算机》2008年第7期48-51,共4页Microelectronics & Computer

基　　金：国家“八六三”计划项目(2006AA10Z237);国家科技支撑计划项目(2006BAD10A1410)

摘　　要：利用隐含语义索引技术设计了一个问答系统,在系统中利用隐含语义索引理论进行查询问题和数据库中的候选问题的相似度计算.主要是通过构造一个语义矩阵,进行奇异值分解消除"噪音"进行实现的.这样更清晰地表示出了词之间的语义相关性,使本系统可以接受被自然语言描述的问题.最后,对整个系统进行实验测试并对测试结果进行了分析,发现本系统比一般的基于VSM等方法实现的系统表现出了明显的优势.A question-answering system is designed using LSI, in which the similarity between query question and question from database is calculated based on the theory＇ of LSI. The arithmetic is realized by structuring a semantic matrix and making a singular value decomposition （SVD） for eliminating noise. So, the correlation among words is shown clearly, which make the system have the capacity of accepting language which is described by natural language. At last, the system is tested. We find that the performance of this system is better than other system based on VSM according to the result of experiment analysis.

关 键 词：隐含语义索引 问答系统 语义计算 向量空间模型 相似度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]