Language Model and Clustering based Information Retrieval
Irene Giakoumi, Christos H. Makris, Yiannis Plegas · 2015
In this paper, we describe two novel frameworks for improving search results. Both of them organize relevant documents into clusters utilizing a new soft clustering method and language models. The first framework is query-independent and takes into account only the inter-document lexical or semantic similarities in order to form clusters. Also, we try to locate the duplicated content inside the formed clusters. The second framework is query-dependent and uses a query expansion technique for the cluster formation. The experimental evaluation demonstrates that the proposed method performs well in the majority of the results.