| Build an index using Krovetz stemmer and stop words removal, taking the two TREC files (with nearly 20,000 documents) given from NY Times. Use "Bush Kerry debates" as the query and perform retrieval using the following models: (1) TFIDF, (2) Okapi, (3) KL, (4) TFIDF with feedback, (5) Okapi with feedback, and (6) KL with feedback. For each of the feedback-based models, use the result set generated by their respective model that used no feedback. With the original models without feedback, retrieve up to 100 documents. With the feedback-based models, retrieve up to 100 documents as results, with the top 10 documents as feedback documents and the top 20 terms as feedback terms. Send these six result-sets to the instructor with "INLS-490: Assignment-5" in the subject. Make sure they are named/labeled properly. (6 points) |
Chirag Shah | Last update: December 19, 2008 |