This chapter presents a Machine Learning (ML)-based framework for predicting energy savings resulting from various Energy Efficiency (EE) renovations. The framework employs three widely recognized tree-based ensemble algorithms: Random Forest (RF), XGBoost, and LightGBM. Predictions generated by these models are aggregated through an ensembling layer with equal weights to mitigate uncertainty and enhance forecasting accuracy. The performance of both the individual models and the ensemble is assessed using a comprehensive database of EE renovation investments. The findings indicate that the ensemble model consistently outperforms the individual ML models across four accuracy metrics. Furthermore, the analysis demonstrates that certain types of EE measures exhibit greater predictability than others, providing valuable insights for stakeholders evaluating investment risks. Importantly, none of the EE categories displayed significantly poor predictive performance.