基于潜在语义分析的跨语言查询扩展方法

Expansion Method for Language-crossed Query Based on Latent Semantic Analysis

针对传统查询扩展方法存在的问题,提出一种基于潜在语义分析的跨语言扩展方法,利用聚类提高扩展文本集合的精度,并用潜在语义分析实现无需翻译的查询扩展,减轻翻译歧义带来的影响。实验结果表明,该方法能够获得较好的性能。

Aiming at the problems existed in traditional method of query expansion, an expansion method for language-crossed query based on Latent Semantic Analysis（LSA） is proposed, which uses clustering to improve accuracy of expansion word sets, and uses LSA to implement query expansion without translation. The impact of error in translation is eliminated. Experimental results show this method achieves better performance.