A Randomised Subspace Gauss-Newton Method for Nonlinear Least-Squares.
Coralia CartisJaroslav M. FowkesZhen ShaoPublished in: CoRR (2022)
Keyphrases
- newton method
- trust region
- convergence analysis
- linear equations
- variational inequalities
- regularized least squares
- quadratic programming
- optimality conditions
- global convergence
- principal component analysis
- high dimensional data
- low dimensional
- linear svm
- nonnegative matrix factorization
- high dimensional
- feature space
- nonlinear programming
- sensitivity analysis
- optimization methods
- feature extraction