SP500215 Automatic Routing and Ad-hoc Retrieval Using SMART: TREC 2 chapter C. Buckley J. Allan G. Salton National Institute of Standards and Technology D. K. Harman X.Y A.B.C R-prec Total Rel recall-prec 1. no fdbk 0.0 8.0.0 3382 6509 2869 2. no expand 0.0 8.8.4 3531 6849 3087 3. little expand 20.10 8.8.4 3756 7192 3345 4. CrR1Ri 300.50 8.16.4 4273 7764 3952 5. crulCi varies varies 4367 7808 4091 Table 6: Routing evaluation For TREC 2, this failure is ameliorated by the base level performance of the global run. If the correct weights are used, the effectiveness of automatic indexing is extremely good. Automatic massive query expansion proves to be very effective for routing. Conven- tional relevance feedback techniques are used to weight the expanded queries. Parameters for the relevance feedback algorithms are esti- mated both over all the queries and for each query individually. The individual query esti- mation perform better (3-4%) but by an in- sufficient amount to be convincing. References [1) Chris Buckley, Gerard Salton, and James Allan. Automatic retrieval with locality information using SMART. In D. K. Har- man, editor, Proceedings of [OCRerr]he Firsi Tex[OCRerr] REtrieval Conference (TREC-1), pages 59-72. NIST Special Publication 500-207, March 1993. [2] James P. Callan and W. Bruce Croft. An evaluation of query processing strate- gies using the tipster collection. In Pro- ceedings of the Sixteenth Annual Inter- national ACM SIGIR Conference on Re- search and Development in Information Retrieval, pages 347-355, June 1993. [3] Norbert Fuhr. Models for retrieval with probabilistic indexing. Information Pro- cessing and Management, 25(1):55-72, 1989. [4] Norbert Fuhr and Chris Buckley. Au- tomatic structuring of text files. ACM Transactions on Information Systems, 9(3):223-248, 1991. [5] Norbert Fuhr and Chris Buckley. Op- timizing document indexing and search 54 term weighting based on probabilistic models. In D. K. Harman, editor, Pro- ceedings of the First Text REtrieval Con- ference (TREC-1), pages 89-99. NIST Special Publication 500-207, March 1993. [6] D. Harman. Towards interactive query expansion. In Proceedings of the Eleventh Jnternational Conference on Research and Development [OCRerr]n Information Re- trieval, pages 321-331. Association for Computing Machinery, 1988. [7] Marti A. Hearst and Christian Plaunt. Subtopic structuring for full-length doc- ument access. In Proceedings of the Six- teenth Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 59-68, June 1993. [8] E. Ide. New experiments in rele- vance feedback. In Gerard Salton, ed- itor, The SMART Retrieval System- Experiments in Automatic Document Processing, chapter 16. Prentice Hall, En- glewood Cliffs, NJ, 1971. [9] J .J. Rocchio. Relevance feedback in in- formation retrieval. In Gerard Salton, editor, The SMART Retrieval System- Experiments in Automatic Document Processing. Prentice Hall, Englewood Cliffs, NJ, 1971. [10] Gerard Salton. Automatic Text Process- ing - the Transformation, Analysis and Retrieval of Information by Computer. Addison-Wesley Publishing Co., Reading, MA, 1989. [11] Gerard Salton. Developments in auto- matic text retrieval. Science, 253:974- 980, August 1991.