A combined component approach for finding collection-adapted ranking functions based on genetic programming.
Humberto Mossri de AlmeidaMarcos André GonçalvesMarco CristoPável CaladoPublished in: SIGIR (2007)
Keyphrases
- ranking functions
- genetic programming
- learning to rank
- document retrieval
- web search
- web search engines
- ranking algorithm
- supervised learning
- evolutionary algorithm
- scoring function
- symbolic regression
- preference learning
- document collections
- beam search
- genetic algorithm
- relevance judgments
- ranking models
- query dependent
- distributed information retrieval
- learning algorithm
- data sets
- collection selection
- directly optimize