Abstract:

In order to facilitate the browse of the search results obtained by search engines and to rapidly and effectively find valuable Web documents, this paper proposes a new clustering algorithm of Web search results based on the conceptual grouping. In this algorithm, first, the co-occurrence networks of characteristic terms are built. Next, the semantic relationships among characteristic terms are mined via the conceptual grouping to form different clusters related to the query topic. Then, the distances between the Web documents and the formed clusters are calculated for the clustering of Web search results. Finally, the cluster labels are selected according to the importance of characteristic terms in the search .results and the clusters. It is indicated by experiments that the proposed algorithm performs better than the k-means algorithm, and that the labels selected by the algorithm are apprehensible.

Key words: information retrieval, search engine, Web document, clustering, conceptual grouping