Rutgers Filtering Work at TREC 2002: Adaptive and Batch
نویسندگان
چکیده
This year at TREC 2002 we participated in the adaptive filtering sub-task of the filtering track with some models for training a Rocchio classifier. Results were poorer than average on the utility type measures. Using simple feature selection produced better than average results on an F-type measure. The key to our approach was the use of pseudojudgments, and an approach to threshold updating. We also participated in the batch filtering sub-task of the filtering track and investigated the use of rank based feature selection techniques in conjunction with a very simple classification rule.
منابع مشابه
The TREC 2002 Filtering Track Report
The TREC–11 filtering track measures the ability of systems to build persistent user profiles which successfully separate relevant and non-relevant documents in an incoming stream. It consists of three major subtasks: adaptive filtering, batch filtering, and routing. In adaptive filtering, the system begins with only a topic statement and a small number of positive examples, and must learn a be...
متن کاملUB at TREC 11: Batch and Adaptive Filtering
This is the first time we participated in TREC filtering track. We submitted four runs: two for adaptive filtering, and two for batching filtering. And these runs come from two separate efforts with very different approaches. One effort treats the filtering problems as standard text categorization problems and solves them using Support Vector Machines (SVM). The second effort is a Language Mode...
متن کاملTREC 11 Experiments at CAS-ICT: Filtering and Web
CAS-ICT took part in the TREC conference for the second time this year and we undertook two tracks of TREC-11. For filtering track, we have submitted results of all three subtasks. In adaptive filtering, we paid more attention to undetermined documents processing, profile building and adaptation. In batch filtering and routing, a centroid-based classifier is used with preprocessed samples. For ...
متن کاملIRIT at TREC 2002: Filtering Track
The experiments we undertaken this year for TREC2002 Filtering track, are focussed on threshold calibration. We proposed a new approach to calibrate the dissemination threshold in an adaptive information filtering. It consists of optimizing a utility function represented by a linearized form of the probability distributions of the scores of the relevant and the non-relevant documents already fi...
متن کاملMicrosoft Cambridge at TREC 2002: Filtering Track
Six runs were submitted for the Adaptive Filtering track, four on the adaptive filtering task (ok11af??), and two on the routing task (msPUM?). The adaptive filtering system has been somewhat modified from the one used for TREC–10, largely for efficiency and flexibility reasons; the basic filtering algorithms remain similar to those used in recent TRECs. For the routing task, a completely new s...
متن کامل