Brain Image Classification Using Optimized Extreme Gradient Boosting Ensemble Classifier

The abnormal development of brain cells is termed as brain tumor that can cause various neurological disorders. The in time detection of tumors can save the life of a person suffering from this dangerous disease. Various imaging techniques are being used to visualize the present condition of the brain so that the treatment will be followed accordingly. Magnetic resonance imaging (MRI) is considered one of the most utilized biomedical imaging techniques. After getting such images of the brain, the next task is the detection of the tumor. The automation in this problem field using machine learning algorithms leads to faster detection in comparison to manual observation. For this purpose, we have used the extreme gradient boosting-based ensemble classifier for brain MRI image classification. The classifier is well optimized by varying the inherent parameters of the classifier and the best score is observed with 96.1% classification accuracy. The training and validation losses are also decreased and recorded as 0.0069 and 0.1214 with proper parameter tuning.