We propose to learn generalized additive models for classification which represents the classifier using a sum of piecewise linear functions and show that a recently proposed fast linear SVM training method (Pegasos) can be adapted to train such models with the same convergence rates. To be able to learn functions on combination of dimensions, we explore the use of random projection features wh...

In this paper, we propose a method for finding certain eigenvalues of a generalized eigenvalue problem that lie in a given domain of the complex plane. The proposed method projects the matrix pencil onto a subspace associated with the eigenvalues that are located in the domain via numerical integration. The projection produces a small pencil with Hankel matrices.

We introduce a new class of robust estimators for generalized linear models which is an extension of the class of projection estimators for linear regression. These projection estimators are defined using an initial robust estimator for a generalized linear model with only one unknown parameter. We found a bound for the maximum asymptotic bias of the projection estimator caused by a fraction ε ...

We introduce the concept of generalized projection operators, i.e., projection integrals over a body in R that generalize the usual result of projected area in a given direction by taking into account shadowing and scattering effects as well as additional convolution functions in the integral. Such operators arise naturally in connection with various observation instruments and data types. We r...

We generalize the grid–projection method for the construction of quasiperiodic tilings. A rather general fundamental domain of the associated higher dimensional lattice is used for the construction of the acceptance region. The arbitrariness of the fundamental domain allows for a choice which obeys all the symmetries of the lattice, which is important for the construction of tilings with a give...

Two projected methods have been proposed to deal with the two main diiculties of the numerical solution of incompressible Stokes problems. While one method enforces the null{ divergency condition through a projection 1], the other 4] projects the solution to satisfy boundary conditions. After summarizing these two techniques, the present report shows how to combine both into an accurate, eecien...

This article explores the SAR back-projection algorithm for stripmap image formation and its characteristics. The article provides a derivation of generalized time-domain back-projection from first principles. It shows that back-projection may be considered an ideal matched filter for SAR. The article presents an analysis of the sensitivity of back-projection to its geometric parameters as well...