An approach for Breast Cancer classification using Neural Networks

Abstract

- Breast Cancer,an increasing predominant death causing disease among women has become a social concern. Early detection and efficient treatment helps to reduce the breastcancerrisk.AdaptiveResonanceTheory(ART1),anunsupervised neural network has become an efficient tool in the classification of breast cancer as Benign(non dangerous tumour) or Malignant (dangerous tumour). 400 instances were pre processed to convert real data into binary data and the classification was carried out using ART1 network. The results of the classified data and the physician diagnosed data were compared and the standard performance measures accuracy, sensitivity and specificity were computed. The results show that the simulation results are analogous to the clinical results.