Nonlinear, statistical data-analysis for the optimal construction of neural-network inputs with the concept of a mutual information.
Frank Heister, Gregor Schock · 2000
Abstract. In this article we focus on a statistical method for nonlinear time series analysis of data-sets used in supervised neural netw ork train-ing. A new method for identifying a minimal neural input-vector with maximum information content is proposed. Further, we demonstrate the capabilit y of the mutual information for nonlinear time series analysis of real measurement data. F rom the viewpoint of information theory this approach provides optimal solutions for a large variety of problems. 1.