Artificial Intelligence based Ensemble Learning Approach for Brain Tumour Classification Using MRI Images

Brain tumours, produced by malignant alterations in brain tissues, are high-risk disorders that require early identification for treatment and prognosis. Current network models perform differently in categorization, and a single model often fails across various evaluation measures. TreerNet, an ensemble learning-based AI classification model, is proposed in this research. It combines TransFG, ResNet50, EfficientNetB4, EfficientNetB7, and ResNeXt101 using ensemble learning weighted averaging. The binary, tertiary, and quaternary brain tumour MRI image classification algorithm is trained using public datasets. Test results reveal that the Treer-Net model has 99.15% accuracy, 99.16% precision, $\mathbf{9 9. 1 5 \%}$ recall, and $\mathbf{9 9. 8 7 \%}$ AUC on the brain tumour three-class dataset. Comparative analysis shows that the suggested Artificial intelligence-based Ensemble Learning method is accurate and efficient, making it ideal for brain tumour identification.