Trust Regions and Ridge Analysis
نویسنده
چکیده
Quasi-Newton methods for numerical optimization exploit quadratic Taylor polynomial models of the objective function. Trust regions are widely used to ensure the global convergence of these methods. Analogously, response surface methods for stochastic optimization exploit linear and quadratic regression models of the objective function. Ridge analysis is widely used to safeguard the optimization of the final quadratic model. This paper describes an intimate connection between trust region methods and ridge analysis, then argues that this connection can be exploited to develop a quasi-Newton approach to stochastic optimization.
منابع مشابه
Analysis of structural characteristics of social capital for increasing the resilience of social- ecological systems (Case study: Kalateh Rudbar, Semnan province)
Researchers studying on adaptation and resilience of systems mainly seek to reduce the vulnerability of extremely complex social-ecological systems against potential changes. Since major portion of natural resources of Iran is interrelated with various users and social groups. Present study focused upon the necessity of studying structural characteristics of social capital of rangeland users in...
متن کاملA generative model and a generalized trust region Newton method for noise reduction
In practical applications related to, for instance, machine learning, data mining and pattern recognition, one is commonly dealing with noisy data lying near some lowdimensional manifold. A well-established tool for extracting the intrinsically low-dimensional structure from such data is principal component analysis (PCA). Due to the inherent limitations of this linear method, its extensions to...
متن کاملConcentration Reversals in Ridge Regression
Ridge regression is often the method of choice in ill–conditioned systems. A canonical form identifies regions in the parameter space where Ordinary Least Squares (OLS) is problematic. A curious but unrecognized property of ridge solutions emerges: Under spherical errors with or without moments, the relative concentrations of the canonical estimators reverse as the ridge scalar evolves, the est...
متن کاملEstimating process capability indices using ridge regression
Process capability indices show the ability of a process to produce products according to the pre-specified requirements. Since final quality characteristics of a product are usually interrelated to its previous amounts in earlier workstations, one need to model and consider the relationship among them to assess the process ca-pability properly. Hence, conducting process capability analysis in ...
متن کاملPrincipal Component Analysis for Minutiae Verification on Fingerprint Image
Minutiae are the two most prominent and well-accepted classes of fingerprint features arising from local ridge discontinuities: ridge endings and ridge bifurcations. However, the preprocessing stage doesn’t eliminate all possible defects in the original gray-level image and the orientation estimation in a poor image is extremely unreliable. In order to further eliminate the false minutiae cause...
متن کامل