Predicting brain tumor presence using machine learning models

A brain tumor is an abnormal growth of the cells within the brain either benign or malignant. The benign ones have relatively slow growth and do not contain any cancerous component, whereas the malignant ones are dangerous as they contain cancerous tendencies and grow at very high rates, extending into the surrounding tissue of the brain. The signs and symptoms of brain tumors depend on position and size but may involve seizures, headaches, abnormalities in vision or hearing, problems in cognition, and inefficient body movements. The diagnosis is usually made by imaging studies such as MRI or CT scan, and the management may be chemotherapy, surgery, radiation therapy, or a combination. The hallmark of improved prognosis in brain tumors is early diagnosis and treatment. The prediction of brain tumors using Machine Learning models like the Extreme Learning Model (ELM) and Decision Tree Classification is done by strategically implementing 2 sophisticated optimizers known as the Mayfly Optimizer and Flying Fox Optimization algorithm. These have contributed to increasing the inherent accuracy of the models. In turn, this leads to the emergence of these models, together with the optimizers, into a synergistic whole that causes a remarkable increase in the accuracy of prediction. In particular, during training, the DTFF model is far ahead among its competitors, providing very impressive accuracy of 0.951, unparalleled by any other model. Then comes the DTMO model, being the runner-up, with an accuracy of 0.932. Conversely, the ELM model, with an accuracy of 0.824, falls short of demonstrating robust predictive capabilities, underscoring its limitations within the prediction process.