LP-SVR Model Selection Using an Inexact Globalized Quasi-Newton Strategy
نویسندگان
چکیده
In this paper we study the problem of model selection for a linear programming-based support vector machine for regression. We propose generalized method that is based on a quasi-Newton method that uses a globalization strategy and an inexact computation of first order information. We explore the case of two-class, multi-class, and regression problems. Simulation results among standard datasets suggest that the algorithm achieves insignificant variability when measuring residual statistical properties.
منابع مشابه
A nonlinear least squares quasi-Newton strategy for LP-SVR hyper-parameters selection
This paper studies the problem of hyperparameters selection for a linear programming-based support vector machine for regression (LP-SVR). The proposed model is a generalized method that minimizes a linear-least squares problem using a globalization strategy, inexact computation of first order information, and an existing analytical method for estimating the initial point in the hyper-parameter...
متن کاملInexact Newton Methods Applied to Under – Determined Systems by Joseph P . Simonis A Dissertation
Consider an under-determined system of nonlinear equations F (x) = 0, F : IR → IR, where F is continuously differentiable and m > n. This system appears in a variety of applications, including parameter–dependent systems, dynamical systems with periodic solutions, and nonlinear eigenvalue problems. Robust, efficient numerical methods are often required for the solution of this system. Newton’s ...
متن کاملInexact-hessian-vector Products for Efficient Reduced-space Pde-constrained Optimization
We investigate reduced-space Newton-Krylov (NK) algorithms for engineering parameter optimization problems constrained by partial-differential equations. We review reduced-space and full-space optimization algorithms, and we show that the efficiency of the reduced-space strategy can be improved significantly with inexact-Hessianvector products computed using approximate second-order adjoints. R...
متن کاملA globally convergent inexact Newton method with a new choice for the forcing term
In inexact Newton methods for solving nonlinear systems of equations, an approximation to the step sk of the Newton’s system J (xk)s =−F(xk) is found. This means that sk must satisfy a condition like ‖F(xk) + J (xk)sk‖ ≤ ηk‖F(xk)‖ for a forcing term ηk ∈ [0,1). Possible choices for ηk have already been presented. In this work, a new choice for ηk is proposed. The method is globalized using a ro...
متن کاملGlobal convergence of an inexact interior-point method for convex quadratic symmetric cone programming
In this paper, we propose a feasible interior-point method for convex quadratic programming over symmetric cones. The proposed algorithm relaxes the accuracy requirements in the solution of the Newton equation system, by using an inexact Newton direction. Furthermore, we obtain an acceptable level of error in the inexact algorithm on convex quadratic symmetric cone programmin...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2013