Geometric Gaussian-Kernel Bolstered Error Estimation for Linear Classification
31 Jan 2013
This is fast implementation of bolstered error estimation algorithm for linear SVM classification.
function [svm_w, svm_b] = svmcoefs(svm_struct)
% Extract hyperplane coefficients from SVM structure
% svm_struct - structure containing information about the trained linear support vector machine
% svm_w - hyperplane coefficients
% svm_b - hyperplane intercept
% NOTE: Works only if the svm was trained with option 'Autoscale', 'off'
sv = svm_struct.SupportVectors;
alphaHat = svm_struct.Alpha;
scaleData = svm_struct.ScaleData;
svm_w = (alphaHat)'*sv;
svm_b = svm_struct.Bias;