THUIR at TREC 2008: Relevance Feedback Track

Tsinghua University Information Retrieval Group (THUIR) has participated into the first Relevance Feedback Track of TREC2008. The TMiner search engine has been used as our text retrieval system, because the processing capability and flexibility of this system on large text data has been testified during many years’ Web Track and Terabyte Track. In the track, we studied two approaches: 1) query expansion, 2) search result re-ranking based on document relevance model. Query Expansion: Terms in the annotated documents (feedback) are used to expand the original query; the new born queries are sent to the search engine for further information retrieval; users get the documents retrieved by the expanded queries. Search Result Re-ranking: The relevance between the annotated documents and other documents are used to influence the search results; users finally get the re-ranked document list. In detail, we have experimented two different methods on which search result re-ranking based: a) Clustering; b) Documents Relevance Model. The rest of this paper is structured as follows. After the introduction of Query Expansion approach in Section 2, Search Result Re-ranking is discussed in Section 3. The evaluation results of the submitted runs are illustrated in Section 4. The last section contains summaries and outlines promising future work.