SDD算法在中文农业网页信息检索的应用

SDD（半离散矩阵分解）算法是潜在语义索引（LSI）的最新技术,弥补了传统SVD算法无法大规模应用的局限,具有压缩比大,响应时间短等优点.利用com技术实现了检索引擎的封装使得服务器端的维护变得轻松,有利于二次开发,同时能够作为一个对象嵌入ASP页面实现了B/S模式,降低了用户使用的门槛.由于SDD算法与具体的语言无关,因此应用面广,在中文领域,尤其是中文农业网页信息检索方面具有实用价值.本文在Microsoft Visual C＋＋.NET开发平台上开发完成了SDD com组件,针对国家农业科学数据中心的10个主要大库共计十万余网页,分别建立索引文件,提供对外检索服务.实践证明SDD算法在中文农业网页检索上具有响应时间快,精度高等优点,是完全切实可行的.

The chinese agricultural web page information retrieval system based on SDD

Semi-discrete decomposition of matrix is the latest technology of LSI,which improves the SVD and can be use in large scale.It has great compress rate,short response time.By using com technology we can encapsulate the search engi ne,so that maintain and second development of server is easy.At the meanwhile S DD com can be used in ASP web page as an object,therefore the B/S model can be used and is convenient for user.Because the SDD has nothing with Language,it ca n be used in many field,especially in Chinese Information Retrieval.We develope d an SDD com under Microsoft visual C++.Net platform then made different index f iles according to ten project of Chinese Agricultural Sciences Data Center,whic h contains more than 10 000 html files to provide web services.In practice SDD has very good performance in Chinese agricultural web page information retrieval.