Minimize a multivariate function, minimize.

This matlab function finds a (local) minimum of a (nonlinear) multivariate function. The user must supply a function which returns the value and partial derivatives wrt all variables. The function uses conjugate gradients and approximate linesearches based on polynomial interpolation with Wolfe-Powel conditions.

Gaussian Processes for Regression, gp.

A bunch of programs for regression using Gaussian Processes with a variety of different parameterised covariance functions.

A simple Kernel Density Estimator, kde.

This matlab function computes a kernel density estimator from a set of examples, by placing Gaussian kernels (with identical masses) on each training data point and adjusting the "widths" to maximize the average leave-one-out log density. In multivariate data, either isotropic or diagonal covariance matrix Gaussians may be used.