Binary Information Press
Search result diversification has recently gained attention as a means to tackle ambiguous queries. Subtopic mining is a basic component of diversified document ranking which will return a selectively ranked list of web pages by given a query. The aim of subtopic mining is to return a ranked list of possible "Subtopic strings" that correspond to a query. In this paper, the authors present an approach for subtopics mining for diversified IR, and to further improve result diversity, they employ a voting method to rank candidate subtopics and a semantic resource HowNet was used to merge those candidate subtopics which may impact diversity.