Bagging gradient-boosted trees for high precision, low variance ranking models.
Yasser GanjisaffarRich CaruanaCristina Videira LopesPublished in: SIGIR (2011)
Keyphrases
- high precision
- low variance
- decision trees
- ranking models
- random forests
- naive bayes
- decision stumps
- high recall
- learning to rank
- real valued
- ranking functions
- ensemble methods
- web search
- ranking algorithm
- benchmark datasets
- loss function
- directly optimize
- cost sensitive
- achieve high precision
- bias variance
- high accuracy
- training set
- information retrieval
- evaluation measures
- feature space
- search engine
- machine learning
- precision and recall
- co occurrence
- pairwise
- training data