Abstract: This paper presents a model of cross-language information retrieval(CLIR) based on language concept space(LCS), the model takes formalized context framework which is formed in language concept space as mediated interface, the framework is three-coordinated and composed of domain, situation and background. The model can avoid the disadvantage of the word sense ambiguity which occurs in the traditional CLIR. The experiments indicate that the LCS based CLIR system has good performance.

Key words: Cross-language information retrieval, Language concept space, Sentences group unit framework

摘要： 提出了一种基于语言概念空间的跨语种信息检索模型，该模型以建立在语言概念空间中的形式化语境单元框架表示处理所需的中间信息，通过用以描述语境单元框架的语义符号间的匹配和生成机制来实现文本的跨语种检索，有助于避开用形式多样的具体语言作为处理中介存在的模糊消解问题。实验证明，这一模型显著改善了检索系统的性能。

关键词: 跨语种信息检索, 语言概念空间, 语境单元框架

CLC Number: 

• TP391