Multi-Merge Budget Maintenance for Stochastic Gradient Descent SVM Training.
Sahar QaadanTobias GlasmachersPublished in: CoRR (2018)
Keyphrases
- stochastic gradient descent
- svm training
- least squares
- loss function
- matrix factorization
- step size
- random forests
- generalization ability
- support vector machine
- regularization parameter
- online algorithms
- multiple kernel learning
- linear svm
- logistic regression
- benchmark datasets
- kernel function
- missing data
- learning algorithm
- weight vector
- importance sampling
- optical flow
- support vector