Building the concept semantic space for large text database.

To overcome such shortcomings of keyword-based systems on the large text database as low efficiency and recall of searching, this paper proposes a novel concept semantic space to describe the large scale of text database efficiently. The proposed concept semantic space describes the text database from m ultiple semantic granularities (i.e. keyword, concept, etc.) and m ultiple semantic dimensions (i.e. association relations, similar relations, etc.), which provides a macroscopic and dynamic v iew of the text database. With the support of concept semantic space, some novel systems can be constructed on the text database to provide novel, efficient and flexible services. In this paper, take association relation for example, the main steps of b u ild in g such a concept semantic space on text database are discussed in detail. In the end, both the experimental results and a prototype system, named as Knowle, show that the proposed concept semantic space is efficient in organizing the text database. [ABSTRACT FROM AUTHOR]