A NOVEL NOISE ROBUST 4TH-ORDER CUMULANTS COST FUNCTION

A novel robust fourth-order cumulants cost function was introduced to enhance the fitting to an underlying function in small data sets with high noise level of Gaussian distribution because higher-order statist ics provide a unique feature of suppressing Gaussian noise processes o f unknown spectral characteristics. The proposed cost function was val idated on the prediction of benchmark sunspot data and an excellent re sult was obtained. The proposed cost function enables the network to p rovide a very low training error and an excellent generalization prope rty, Our result indicates that the network trained by the proposed cos t function can, at most, provide 74% reduction of the normalized test error in the benchmark test.