Gliomas are among the most common and lethal types of brain tumors, with significant implications for patient prognosis and treatment strategies. Accurate and transparent classification of glioma grades is crucial for effective clinical decision-making. This study aims to empower glioma prognosis by employing a robust ensemble of machine learning algorithms and deep learning techniques, augmented with interpretative insights derived from Explainable AI (XAI) strategies. In this research, we classify glioma grades into two categories: Low-Grade Gliomas (LGG) and Glioblastoma Multiforme (GBM). The classification models implemented include Random Forest, Decision Trees, Logistic Regression, K-Nearest Neighbors (KNN), Adaboost, Support Vector Machine (SVM), CatBoost, LightGBM (LGBM), and XGBoost, alongside deep learning models such as Artificial Neural Networks (ANN) and Convolutional Neural Networks (CNN).  To enhance the interpretability of these models, we integrate four alternative XAI strategies, focusing on the SHapley Additive exPlanations (SHAP) algorithm. SHAP provides a detailed understanding of feature importance and model predictions, ensuring that the classification process remains transparent and comprehensible to clinicians. Through comprehensive experimentation and evaluation, this study demonstrates the potential of combining advanced machine learning techniques with XAI to improve glioma grade classification accuracy while maintaining model transparency. The findings underscore the importance of explainability in clinical applications, fostering trust and facilitating the adoption of AI-driven solutions in medical practice. This approach not only aids in precise glioma prognosis but also sets a benchmark for future research in the intersection of machine learning and healthcare.

Keywords: Glioma prognosis, machine learning, deep learning, explainable AI, SHAP, glioma grade classification, LGG, GBM.