Diagnosis of Breast Cancer Using Novel Hybrid Approaches with Genetic Algorithm

Cancer is a group of diseases which are formed by the uncontrolled proliferation and growth of tissues and cells in organs and their treatment and approach are different from each other. Cancer disease shows too fast metastasis. Therefore, early diagnosis and treatment is very important. Technological advances in computer and electronics, early stages of cancer increased the probability of correct diagnosis. Especially in recent years, better results are obtained in the diagnosis of cancer with the studies based on machine learning. In this study, hybrid approaches were proposed for diagnosis of breast cancer. XGBOOST and Artificial Neural Network (ANN) algorithms were employed by hybridizing with genetic algorithm (GA) to improve classification accuracy. The performance analysis of the proposed approaches was presented with ‘Wisconsin’ dataset taken from UCI machine learning repository. Numerical results showed that the proposed hybrid XGBOOST-GA approach significantly outperformed the classical prediction algorithms and the best classification accuracy was achieved.