Background: With DNA microarray data, selecting a compact subset of discriminative genes from thousands of genes is a critical step for accurate classification of phenotypes for, ...
We investigated the geometrical complexity of several high-dimensional, small sample classification problems and its changes due to two popular feature selection procedures, forw...
Erinija Pranckeviciene, TinKam Ho, Ray L. Somorjai
Support vector machines are trained by solving constrained quadratic optimization problems. This is usually done with an iterative decomposition algorithm operating on a small wor...
This paper presents a new method for the selection of the two hyperparameters of Least Squares Support Vector Machine (LS-SVM) approximators with Gaussian Kernels. The two hyperpar...
Amaury Lendasse, Yongnan Ji, Nima Reyhani, Michel ...
: A novel support vector machine method for classification is presented in this paper. A modified kernel function based on the similarity metric and Riemannian metric is applied ...