Research on keyword information retrieve based on semantic

Abstract

To deal with the problem of semantic missing during keyword retrieval, this paper proposes a semantic-based keyword information retrieval approach. Firstly, an improved word-concept relevance method, which measures the relevance between word and concept by considering both the internal and external correlations, is proposed. Next, by importing the improved word-concept relevance method into the classical statistical language model, a new statistical language model which is based on the relevance of word and concept is proposed. The experiments demonstrated good efficiency and performance of the approach proposed in this paper. © Springer-Verlag Berlin Heidelberg 2011.