Non-convex Regularizations for Feature Selection in Ranking With Sparse SVM.
Léa LaporteRémi FlamaryStéphane CanuSébastien DéjeanJosiane MothePublished in: CoRR (2015)
Keyphrases
- feature selection
- feature ranking
- support vector
- high dimension
- support vector machine
- sparse pca
- group lasso
- minimization problems
- support vector machine svm
- multi class
- ranking svm
- knn
- text categorization
- variable selection
- ranking algorithm
- multiple kernel learning
- selected features
- feature space
- learning to rank
- high dimensional
- feature set
- convex hull
- svm classifier
- classification accuracy
- reduced set
- feature subset
- text classification
- feature extraction
- convex relaxation
- input features
- sparse representation
- naive bayes
- information gain
- ranking functions
- convex optimization
- support vectors
- ensemble classifier
- web search
- bayes classifier
- multi task
- web image annotation
- classification performances
- classification algorithm
- fold cross validation
- model selection
- small sample
- loss function
- dimensionality reduction
- machine learning
- feature vectors
- training data
- kernel function
- cross validation