Classification of Prostate Cancer using Wavelet Neural Network

Abstract

Prostate cancer is the century disease that endanger the life of men. The earlier to diagnose the disease, the probability of curing this disease is higher. Therefore, new approaches of diagnosis is required to effectively detect the prostate cancer in early stage compared to the traditional methods. Therefore, WNN is a new adopted approach in prostate cancer diagnosis. Morlet function is used as an activation function of wavelet neural network (WNN) and back propagation (BP) is applied to train the Wavelet network. WNN classifies prostate cancer according to three factors: patient age, PSA level, and prostate volume. WNN performance is evaluated based on the percentage of classification and the computational complexity of several cases. The results of the simulation show that WNN has lower mean squared error (MSE) than the Neural Network (NN).