How to build high quality L2R training data: Unsupervised compression-based selective sampling for learning to rank.
Rodrigo M. SilvaGuilherme C. M. GomesMário S. AlvimMarcos André GonçalvesPublished in: Inf. Sci. (2022)
Keyphrases
- selective sampling
- learning to rank
- supervised learning
- active learning
- ranking functions
- training data
- active sampling
- random sampling
- unsupervised learning
- training set
- semi supervised
- unlabeled data
- labeled data
- learning algorithm
- training examples
- unlabeled examples
- feature selection
- information retrieval
- ranking svm
- training samples
- semi supervised learning
- class labels
- support vector machine
- generalization error
- learning tasks
- learning problems
- data sets
- loss function
- classification accuracy
- decision trees
- class distribution
- co training
- learning to rank algorithms
- machine learning
- transfer learning
- ranking algorithm
- document retrieval
- naive bayes
- multi class
- upper bound
- learning process
- feature vectors
- similarity measure
- multimedia