Your team is working on a machine learning project with highly imbalanced classes. Which Spark ML algorithm would you choose to tackle the class imbalance and enhance model performance?