Feedback RBF GMDH-Type Neural Network Using Principal Component-Regression Analysis and Its Application to Medical Image Diagnosis of Lung Cancer

Abstract

The feedback Radial Basis Function (RBF) Group Method of Data Handling (GMDH)-type neural network algorithm is applied to the medical image diagnosis of lung cancer. In this feedback RBF GMDH-type neural network algorithm, the principal component-regression analysis is used to protect multi-colinearity which is occurred in the learning calculation of neurons, and the accurate and stable neural network architectures are organized. Furthermore, the structural parameters such as the number of feedback loops, the number of neurons in the hidden layers and the relevant input variables are automatically selected so as to minimize the prediction error criterion defined as Akaike's Information Criterion (AIC) or Prediction Sum of Squares (PSS). This feedback RBF GMDH-type neural network is applied to the medical image diagnosis of lung cancer and the results are compared with those of the conventional neural network trained using the back propagation algorithm. It is shown that the feedback RBF GMDH-type neural network algorithm is useful for the medical image diagnosis of lung cancer since the optimum neural network architecture is automatically organized so as to fit the complexity of the medical images.