Improved Feature Selection and Redundance Computing - THUIR at TREC 2004 Novelty Track.
Liyun RuLe ZhaoMin ZhangShaoping MaPublished in: TREC (2004)
Keyphrases
- feature selection
- text categorization
- information retrieval
- feature set
- mutual information
- blog opinion retrieval
- machine learning
- pseudo relevance feedback
- document retrieval
- test collection
- support vector
- feature extraction
- novelty detection
- model selection
- information gain
- improved algorithm
- feature weighting
- ad hoc retrieval
- expert search
- clef ip
- learning algorithm