Locally Optimized Kernels
نویسندگان
چکیده
Support Vector Machines (SVM’s) with various kernels have become very successful in pattern classification and regression. However, single kernels do not lead to optimal data models. Replacing the input space by a kernel-based feature space in which the linear discrimination problem with margin maximization is solved is a general method that allows for mixing various kernels and adding new types of features. We show here how to generate locally optimized kernels that facilitate multi-resolution and can handle complex data distributions using simpler models than the standard data formulation may provide.
منابع مشابه
Locally Supported Kernels for Spherical Spline Interpolation
By the use of locally supported basis functions for spherical spline interpolation the applicability of this approximation method is spread out since the resulting interpolation matrix is sparse and thus eecient solvers can be used. In this paper we study locally supported kernels in detail. Investigations on the Legendre coeecients allow a characterization of the underlying Hilbert space struc...
متن کاملRoasting Process Optimization of Walnut Kernels for the Preparation of Walnut Cream Using Response Surface Methodology
Roasting has considerable effects on the quality of cream made of nuts. In this study, the roasting conditions of walnut kernels were optimized based on the stability parameters of the produced cream. Temperatures of 100-150°C for 10-30 minutes were used to roast walnut kernels. The amount of oil separation, peroxide, acidity and Thiobarbituric acid values of the cream, as well as color paramet...
متن کاملA Family of High-Performance Matrix Multiplication Algorithms
During the last half-decade, a number of research efforts have centered around developing software for generating automatically tuned matrix multiplication kernels. These include the PHiPAC project and the ATLAS project. The software products of both projects employ brute force to search a parameter space for blockings that accommodate multiple levels of memory hierarchy. We take a different ap...
متن کاملCompiler-Optimized Kernels: An Efficient Alternative to Hand-Coded Inner Kernels
The use of highly optimized inner kernels is of paramount importance for obtaining efficient numerical algorithms. Often, such kernels are created by hand. In this paper, however, we present an alternative way to produce efficient matrix multiplication kernels based on a set of simple codes which can be parameterized at compilation time. Using the resulting kernels we have been able to produce ...
متن کاملSupport Vector Regression Using Mahalanobis Kernels
In our previous work we have shown that Mahalanobis kernels are useful for support vector classifiers both from generalization ability and model selection speed. In this paper we propose using Mahalanobis kernels for function approximation. We determine the covariance matrix for the Mahalanobis kernel using all the training data. Model selection is done by line search. Namely, first the margin ...
متن کامل