PERFORMANCE OF MULTILAYER FEEDFORWARD AND RADIAL BASE FUNCTION NEURALNETWORKS IN CLASSIFICATION AND MODELING

Neural networks have been used in multiple applications, but as a kind of black box for dealing with problems where there is no a priori inf ormation about the data, This means that the model is constructed base d solely upon information obtained from the data themselves. This seem s to be a good property but makes it difficult to validate the models obtained. The classification properties of neural classifiers are usua lly described by the percentage of correctly classified objects in a t est set. Since these straight methods are only based on discrimination , no information can be obtained in a statistical way. In this paper, on a simulated data set, two different types of neural networks, MLF ( multi layer feedforward) and RBF (radial base function), are applied t o solve a classification problem. The modelling ability, stability and reproducibility of this kind of networks are studied based on various different networks independently trained on the same data set with a predetermined value for the sensibility and specificity, Robustness to different kinds of error is also studied by means of Monte Carlo simu lations adding noise at different levels and from different theoretica l distributions. Further to this, an analysis based on principal compo nents is carried out to study the apparently different networks obtain ed. The simulation studies reveal that both types of networks perform well enough to reproduce the input space, For RBF networks, due to the local approach, the study showed some properties related to sensibili ty and specificity which are relevant in practical problems.