Authors: Hiroyuki Seki, Kokoro Nakagawa, Yoshiaki Takata
Tags: 2000, conceptual modeling
A method for supporting WWW retrieval by constructing a flexible category structure adaptable to the user’s search intention is proposed. The method uses categorization viewpoints as a priori knowledge, where a categorization viewpoint is a finite set of consistent category names. A set of documents retrieved by initial keywords is decomposed by categorization viewpoints and each decomposition is scored by clearness or entropy. The user selects an appropriate decomposition by considering the score. The decomposition is recursively performed until a category structure of reasonable size is obtained. Experimental results show that the sets of documents decomposed by the proposed method have higher precision than those decomposed by clustering (K-means). It is also shown that both the scores based on clearness and entropy of the decomposition have relatively high correlation with the precision.Read the full paper here: https://link.springer.com/chapter/10.1007/3-540-45394-6_15
