A Non-Functional Prototype at TREC 2003

As a first attempt at participation in the TREC competition, we built a system which produced some preliminary results, but was unable to generate the quality of results that we expected. While we were able to submit four base-line runs, bugs were discovered in the final hours before the deadline making it impossible to submit results using our intended implementation. We have since found additional coding errors, making our submitted results expectedly poor. The size of our index dataset was approximately 3.8GB without compression. We did not use term position information nor any kind of phrasal indexing.

[1]  Gerard Salton,et al.  Term-Weighting Approaches in Automatic Text Retrieval , 1988, Inf. Process. Manag..