本文提出了一种基于ＳＶＤ（奇异值分解）［１］的双语信息过滤［２］算法，将双语文档进行了统一的表示，使得适应于单语过滤的算法可以方便地用于双语过滤，同时对文档向量进行了压缩，滤去了噪声。在应用方面，将双语过滤算法用于互联网上的个性化主动信息过滤。 In this paper, we propose a bilingual message filtering algorithm [2] based on SVD (singular value decomposition) [1], which presents a unified representation of bilingual documents so that the algorithm adapted to single language filtering can be conveniently used for bilingual filtering. Meanwhile, The document vector was compressed, filtered out of noise. On the application side, the bilingual filtering algorithm is used for personalized active information filtering on the Internet.