On the Use of an Intermediate Class in Boolean Crowdsourced Relevance Annotations for Learning to Rank Comments.
Alberto Barrón-CedeñoGiovanni Da San MartinoSimone FiliceAlessandro MoschittiPublished in: SIGIR (2017)
Keyphrases
- learning to rank
- relevance judgments
- information retrieval
- ranking functions
- learning to rank algorithms
- test collection
- query dependent
- loss function
- evaluation measures
- relevance assessments
- ranking svm
- ranking models
- evaluation metrics
- retrieval systems
- document retrieval
- user feedback
- collaborative filtering
- directly optimize
- learn a ranking function
- direct optimization
- relevance feedback
- metadata
- supervised learning
- clickthrough data
- digital libraries
- retrieval effectiveness
- information retrieval systems
- multi class
- ranking list
- training data
- machine learning