Breast Cancer Detection Using BA-BP Based Neural Networks and Efficient Features

This paper presents an accurate hybrid system for recognizing breast cancer tumours which includes three main modules: feature extraction module, training module and classifier module. In feature extraction module, fuzzy feature has used as an effective classifier input. In training module, a hybrid bees algorithm (BA) back-propagation (BP) algorithm is proposed to train the classifier. This module enjoys the advantages of global search of BA and local search of BP algorithm. In classifier module, multi-layer perceptron (MLP) neural network is used. The proposed system is tested on Wisconsin breast cancer (WBC) database and the simulation results show that the recommended system has high recognition accuracy in comparison with other methods.