Non-integer norm regularization SVM via Legendre-Fenchel duality.
Jian-Wei LiuYuan LiuPublished in: Neurocomputing (2014)
Keyphrases
- norm regularization
- multiple kernel learning
- max margin
- support vector
- support vector machine svm
- convex optimization
- canonical correlation analysis
- loss function
- multi class
- group lasso
- regularization methods
- total variation
- support vector machine
- feature selection
- sparse representation
- matrix completion
- maximum margin
- knn
- kernel function
- svm classifier
- kernel learning
- kernel methods
- semi supervised learning
- markov networks
- support vectors
- machine learning
- pattern recognition
- ls svm
- multi label
- conditional random fields
- linear combination
- stochastic gradient descent
- image processing
- multi task
- structured prediction
- feature vectors
- face recognition
- regression model
- text categorization