BIT at TREC 2010 Blog Track: Faceted Blog Distillation
Abstract
This paper presents the work done for the TREC 2010 faceted blog distillation task. As the approach used in TREC 2009, a mixture of language models based on global representation is employed to rank the entire blogs by relevance and facets. The parameters in our approach are adjusted according to the experimental results in TREC 2009. In addition, we make use of the results evaluated in TREC 2009 to train a SVM classifier. This classifier is used to filter and re-rank the results obtained by the mixture model.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 2010
- Accession Number
- ADA546550
Entities
People
- Chunxia Zhang
- Peng Jiang
- Qing Yang
- Zhendong Niu
Organizations
- Beijing Institute of Technology