中英文双语交叉过滤的逻辑模型

Logic Model for Chinese-English Cross-language Text Filtering

文章简要地描述了文本过滤的背景，提出了基于潜在语义索引的中英文双语交叉过滤的逻辑模型。其基本思想是改进双语交叉过滤中基于词汇对译的方法，而是利用双语文本中潜在的语义结构，作为用户模板与文本匹配的基础。将出现的双语词汇和文本映射为语义空间的向量，不必翻译对译词，甚至不需要出现相应的对译词，也能匹配成功，极大地改善了交叉过滤的精度，效果良好。

This paper briefly describes the background of text filtering and puts forward the logic model for ChineseEnglish cross-language text filtering based on hatent Semantic Indexing. The main idea is showed as follows; It takesthe latent semantic structures as a basis of the matching between user profile and texts instead of the word to wordtranslations,and either texts or words in bi-language can be represented as vectors in new semantic space. As a result,inner production acts as similarity measure,so profiles and texts could match successfully without corresponding wordsin another language. The experiment shows that it can markedly improve the efficiency of text filtering.