Back Propagation Neural Network by Comparing Hidden Neurons: Case study on Breast Cancer Diagnosis

This paper investigates the potential of applying the feed forward neural network architecture for the classification of breast cancer. Back-propagation algorithm is used for training multi-layer artificial neural network. Missing values are replaced with median method before the construction of the network. This paper presents the results of a comparison among ten different hidden neuron initialization methods. The classification results have indicated that the network gave the good diagnostic performance of 99.28%.