Consensual Neural Networks

A new neural network architecture is proposed and applied in classification of data from multiple sources. The new arclhitecture is called a consensual neural network and its relation to hierarchical and ensemble neural networks is discussed. The consenr;ual neural nebwork architecture is based on statistical consensus theory and invol.ves using non-linearly transformed input data. The input data are transformed several times and the different transformed data are applied as if they were independent inputs. The independent inputs are c!lassified using stage neural networks and the 0utput.s from the stage: networks are then weighted and combined to make a decision. Experimental results based on remote sensing data and geographic data are given. The performance of the consensual neur,al network archi.tecture is compared to that of a two-layer conjugate-gradient backpropagation neural network. The results with the proposed neural network architecture compare favourably to the backpropagation method in terms of classification accuracy.