Stochastic Gradient Descent with Barzilai-Borwein update step for SVM.
Krzysztof SopylaPawel DrozdaPublished in: Inf. Sci. (2015)
Keyphrases
- stochastic gradient descent
- loss function
- step size
- least squares
- matrix factorization
- svm training
- support vector machine
- random forests
- svm solvers
- multiple kernel learning
- regularization parameter
- online algorithms
- support vector machine svm
- support vector
- weight vector
- importance sampling
- training data
- convergence rate
- kernel function
- feature selection
- cost sensitive
- binary classification
- support vectors
- svm classifier
- machine learning algorithms
- image restoration
- multi class
- knn