结合本体论和统计方法的跨语言信息检索模型

CLIR model based on a combination of ontology and statistical method

为了更有效地提高跨语言信息检索的性能,结合本体论和统计方法的特性,提出一种混合的跨语言信息检索模型.在该语言模型的结构上,提出一种本体描述框架,构造了一个形式化的语言本体知识表示,通过典型语料学习,融合了语法、语义、句法等多元信息,建立了源语言本体知识库.在跨语言信息检索的实际应用中,利用本体表示,获得初始的检索文档集,再基于源语言本体知识库,对全部候选文档重新排序,以提高TopN排列的精确度.利用NTCIR-3Workshop中的中英文跨语言信息检索数据集对该语言模型进行了评价,相关实验结果表明,该方法取得了较满意的实验效果.

For improving the performance of cross-lingual information retrieval, a hybrid language presented based on a combination of ontology and statistical method. In the structure of the languag model is e model, an ontology description frame was given and a linguistic ontology knowledge presentation was determined. A linguistic ontology knowledge bank of source language was created, which combines with semantic, pragmatic and syntactic by learning typical corpus. In cross-lingual information retrieval, the initial document set will be obtained by ontology presentation and all documents will be re-ordered based on linguistic ontology knowledge of source language for improving the precision of Top N rank. The cross-lingual information retrieval data set in NTCIR-3 Workshop was used to evaluate the performance of the language model. The results indicate that the proposed method improves the precision of nature language processing.