I have a dataset with 2 classes (churners and non-churners) in the ratio 1:4. I used Random Forests algorithm via Spark MLlib. My model is terrible at predicting churn class
