Bayesian Optimized Artificial Neural Network for Breast Cancer Classification

Breast cancer is the most prevalent cause of mortality worldwide, with an estimated million people dying from it every year. One suitable strategy to lower the death rate from breast cancer is early detection. In this context, in the literature, numerous classifiers are reported. However, hyperparameter selection is an issue in most classifiers as it directly controls their performance. Compared to traditional classifiers, artificial neural networks (ANNs) are known to be highly sensitive to the choice of hyperparameters. The efficiency of manual tuning for ANN architectures has improved over time. Nevertheless, hyperparameter optimization remains a significant challenge, especially for new architecture, tasks, or datasets. In the current work, a Bayesian optimized ANN (BOANN) is proposed for the breast cancer classification task. The study utilizes the Bayesian optimization technique to obtain the best hyperparameters set for the chosen ANN model. Two standard datasets involving numerical feature values are utilized to evaluate the robustness of the suggested model. Experimental result shows that the proposed model outperforms the baseline methods used for the breast cancer classification task.