潜在语义索引方法在信息过滤中的应用

摘要

Information Filtering is one of the latest web information technologies,which aims at active information recommendation from web server to clients. How to build and represent the Topic Profile of client users is one of the key technologies in information filtering. In this paper,Latent Semantic Indexing,or LSI,is applied in building the Topic Profile model and filtering a great deal of Chinese technological Documents on the WWW. The preliminary experiment shows that,compared with Vector Space Model,the precision of LSI is greatly improved.%信息过滤是一种WEB信息服务的新技术，旨在实现网络服务器向客户端主动的信息推送，其核心技术之一是用户兴趣主题模型的表示。文章利用潜在语义索引方法构建用户兴趣主题模型，并对网上大量的中文科技文献信息进行过滤。初步实验的结果表明，该方法与传统的向量空间方法相比，效率有明显提高。