An almost cyclic 2-coordinate descent method for singly linearly constrained problems.
Andrea CristofariPublished in: Comput. Optim. Appl. (2019)
Keyphrases
- constrained problems
- coordinate descent method
- simple and easy to implement
- constraint propagation
- convex optimization
- constraint satisfaction
- loss function
- linear support vector machines
- optimization problems
- linear svm
- multi class boosting
- nonnegative matrix factorization
- logistic regression
- machine learning
- search algorithm
- data representation
- stochastic local search
- neural network
- support vector
- training data
- constraint satisfaction problems
- np hard