Feature extraction from stochastic process samples
Abstract
To analyse a stochastic process described by samples drawn from different classes, a method for automatic extraction of discriminant features in reduced dimension space is proposed. To be effective, dimension reduction should be achieved with minimum loss of information. The proposed method is based on the search for an optimal regression between representation space and feature space according to class information. Information is measured using a mutual information estimate. A nonparametric entropy estimate and a stochastic distributed optimisation algorithm are used to solve this problem. An experimental study of simulated problems shows the efficiency of the proposed method.