A support vector method for optimizing average precision
- 23 July 2007
- proceedings article
- Published by Association for Computing Machinery (ACM)
- p. 271-278
- https://doi.org/10.1145/1277741.1277790
Abstract
Machine learning is commonly used to improve ranked retrieval systems. Due to computational difficulties, few learning techniques have been developed to directly optimize for mean average precision (MAP), despite its widespread use in evaluating such systems. Existing approaches optimizing MAP either do not find a globally optimal solution, or are computationally expensive. In contrast, we present a general SVM learning algorithm that efficiently finds a globally optimal solution to a straightforward relaxation of MAP. We evaluate our approach using the TREC 9 and TREC 10 Web Track corpora (WT10g), comparing against SVMs optimized for accuracy and ROCArea. In most cases we show our method to produce statistically significant improvements in MAP scores.Keywords
This publication has 12 references indexed in Scilit:
- Learning a ranking from pairwise preferencesPublished by Association for Computing Machinery (ACM) ,2006
- Adapting ranking SVM to document retrievalPublished by Association for Computing Machinery (ACM) ,2006
- The relationship between Precision-Recall and ROC curvesPublished by Association for Computing Machinery (ACM) ,2006
- A Markov random field model for term dependenciesPublished by Association for Computing Machinery (ACM) ,2005
- Learning to rank using gradient descentPublished by Association for Computing Machinery (ACM) ,2005
- A support vector method for multivariate performance measuresPublished by Association for Computing Machinery (ACM) ,2005
- Ensemble selection from libraries of modelsPublished by Association for Computing Machinery (ACM) ,2004
- Optimising area under the ROC curve using gradient descentPublished by Association for Computing Machinery (ACM) ,2004
- Document language models, query models, and risk minimization for information retrievalPublished by Association for Computing Machinery (ACM) ,2001
- THE PROBABILITY RANKING PRINCIPLE IN IRJournal of Documentation, 1977