Modeling Depression Data: Feed Forward Neural Network vs. Radial Basis Function Neural Network

Abstract

Depression is a serious affliction that affects a large fraction of the global populace. Due to the widely varying symptoms the diagnosis poses a unique problem based on uncertainty. This paper proposes an approach to tackle the aspect of uncertainty using Soft Computing techniques, which are trained using real life medical data. We have developed two forms of intelligent Neural Network models to help in obtaining a reasonably accurate diagnosis. Trials with test data have yielded nominal Mean Squared Error. Hence this could prove to be a useful tool in automated diagnosis of depression.