Deep Learning based on Classification of Breast Cancer Diagnosis using Binary Grey Wolf Optimization Algorithm

For breast cancer treatment to be effective, early detection is key. Though current methods encounter obstacles in attaining ideal accuracy, Computer-Aided Diagnosis systems remain indispensable in the automated processing, interpretation, grading, as well as early identification of breast cancer through mammography images. This research overcomes these shortcomings by combining a Support Vector Machines radiation basis function Kernel with the upgraded binary Grey Wolf Optimizer, which is inspired by quantum mechanics. Finding the best Support Vector Machine features is the goal of this hybrid strategy, which tries to improve breast cancer classification accuracy. The requirement for better categorization performance in comparison to current optimizers like Genetic Algorithm and Particle Swarm Optimisation is what drives this hybridization. Analyse the MIAS dataset to determine how well the suggested BGW method performs in terms of accuracy, sensitivity, and specificity, among other metrics. In addition, we will compare the outcomes after investigating the use of BGWO in feature selection. Utilising a tenfold cross-validation datasets split, the experimental results show that the proposed BGWO method achieves better results than state-of-the-art classification methods using the MIAS dataset. Specifically, the mean accuracy is 99.65%, sensitivity is 98.99%, and specificity is 100%.