论文信息 - The Robert Gordon University's HARD Track Experiments at TREC 2004

The Robert Gordon University's HARD Track Experiments at TREC 2004

The High Accuracy Retrieval from Documents (HARD) track explores methods of improving the accuracy of document retrieval systems. As part of this track, the participants have investigated how information about a searcher’s context can be used to improve retrieval performance [Allan, 2003; Allan, 2004]. Searchers, referred to as assessors in this track, produce TREC-style search topics. Additionally, assessors are asked to specify values from pre-defined categories of metadata, that relate to the context of the search, as opposed to the topic of the search. The categories and values of the metadata used for the HARD track in 2004 are:

[1] Nicholas J. Belkin,et al. Rutgers' HARD and Web Interactive Track Experiments at TREC 2003 , 2003, TREC.

[2] W. Bruce Croft,et al. Predicting query performance , 2002, SIGIR '02.

[3] James Allan,et al. HARD Track Overview in TREC 2003: High Accuracy Retrieval from Documents , 2003, TREC.

[4] Joemon M. Jose,et al. A Retrieval Mechanism for Semi-Structured Photographic Collections , 1997, DEXA.

[5] Ellen M. Voorhees,et al. The Twelfth Text Retrieval Conference, TREC 2003 , 2004 .

[6] Joemon M. Jose,et al. Spatial querying for image retrieval: a user-oriented evaluation , 1998, SIGIR '98.

[7] James Allan,et al. HARD Track Overview in TREC 2004 (Notebook) High Accuracy Retrieval from Documents , 2004 .

[8] Jana Urban,et al. AN ADAPTIVE APPROACH TOWARDS CONTENT-BASED IMAGE RETRIEVAL , 2003 .

[9] Clement T. Yu,et al. Multiple evidence combination in image retrieval: Diogenes searches for people on the Web , 2000, SIGIR '00.