Bio-inspired algorithms for diagnosis of breast cancer

摘要

Most commonly found cancer among women is breast cancer. Roughly 12% of women grow breast cancer during their lifetime. It is the second prominent fatal cancer among women. Breast cancer diagnosis is necessary during its initial phase for the proper treatment of the patients to lead constructive lives for an extensive period. Many different algorithms are introduced to improve the diagnosis of breast cancer, but many have less efficiency. In this work, we have compared different bio-inspired algorithms including artificial bee colony optimisation, particle swarm optimisation, ant colony optimisation and firefly algorithm. The performances on these algorithms have been measured for UCI Dataset of Wisconsin Diagnostic Breast Cancer, and the results have been calculated using different classifiers on the selected features. After the experiment, it is seen that BPSO has shown maximum accuracy of 96.45% and BFA has shown considerable results of 95.81% with six features which is minimum of all algorithms.