The Characteristic Analysis of the DARE Visual Space     

Jin Zhang 《Information Retrieval》2001,4(1):61-78

In this paper a distance-angle-based visual retrieval tool DARE is introduced. The distance-based similarity distribution, angle-based similarity distribution, and the differences of their distributions in the visual space are analyzed. The document cluster analysis in the visual space and the document cluster comparison between the document space and the visual space are addressed. A new concept—Distance to Reference Axis—is introduced to better understand the visual space. The impact of other operations in DARE on the document distribution is discussed. Future research directions including significance of the index term distribution in the visual space and a user study are addressed.  相似文献