SP500215 TREC-2 Routing and Ad-Hoc Retrieval Evaluation using the INQUERY System chapter W. Croft J. Callan J. Broglio National Institute of Standards and Technology D. K. Harman higher in terms of average precision than the best queries that were not modified using relevance judgements). It is becoming apparent that techniques that may not work well in relevance feedback situations with few identified relevant documents, may be very effective in routing where there are many more relevant documents identified. We are currently doing experiments with different forms of weighting, including the use of identified non-relevant documents. With regard to improving the performance of adhoc queries, we are continuing to carry out experiments with different ways of estimating the probabilities (or tf.idf weights) needed for the inference net, and with different forms of paragraph-level matching. Finally, as men- tioned earlier, we have seen some significant improvements using automatic query expansion based on corpus analysis. References [1] J. P. Callan and W.B. Croft. An evaluation of query processing strategies using the TIP- STER collection. In Proceedings of ACM SICIR International Conference on Research and Development in Information Retrieval, pages 347-356, 1993. [2] J. P. Callan, W.B. Croft, and S.M. Harding. The INQUERY retrieval system. In Proceed- ings of the 3rd International Conference on Database and Expert Systems Applications, pages 78-83, 1992. [3] W. Bruce Croft and Howard R. Turtle. Text retrieval and inference. In P. Jacobs, editor, Text-Based Intelligent Systems, pages 127-156. Lawrence Erlbaum, 1992. [4] Hideo Pujii and W. Bruce Croft. A comparison of indexing techniques for japanese text retrieval. In Proceedings of the ACM SIGIR International Conference on Research and Development in Information Retrieval, pages 237-246. ACM, 1993. [5] David Haines and W. Bruce Croft. Relevance feedback and inference networks. In Proceedings of the ACM SIGIR International Conference on Research and Development in Information Retrieval, pages 2-11. ACM, 1993. [6] H.R. Turtle and W.B. Croft. Evaluation of an inference network-based retrieval model. ACM Transactions on Information Systems, 9(3):187-222, 1991. [7] H.R. Turtle and W.B. Croft. A comparison of text retrieval models. Compt£ter Jot£rnal, 1992. To appear. 83