New Multilayer Neural Networks With NO Estimator and Winsorized Mean

Multilayer feed forward neural networks have been widely used for prediction, forecastingand classification over the past few years. However, it is a known fact that the mostlypreferred Mc - Culloch Pitts neuron model used in these network types does not give asuccessful prediction performance in data sets with outliers. Therefore, robust neuron modelsusing median and trimmed mean aggregation functions have been proposed. However, thesestudies were generally focused on time series forecasting. In this study, we developed newneuron models using NO estimator and the Winsorized mean for prediction, classification,and time series forecasting. NO is a quantile estimator with weights determined by using asubsampling approach. For estimating a population quantile, it uses all order statistics in asample and the accompanying weights of the order statistics are calculated from a BinomialDistribution. The proposed NO and Winsorized mean neuron models are not sensitive tooutlying observations. Back propagation, particle swarm optimization and artificial beecolony optimization algorithms were used when training multilayer neural networks andseveral activation functions such as sigmoid, hyperbolic, tangent, and rectified linear unitwere tried. All steps of the study were performed using statistical programming language R.The written functions of the proposed neural networks enable the prediction of newobservations and observing the change of errors at each iteration by providing a dynamic plot.The developed methods were applied on the real data sets and their performances werecompared with the existing ones. More successful results were achieved in terms of differentperformance criterions