How to decide the best machine learning model according to accuracy, precision, recall, f1 score and roc score?

See more:

machine-learning

Which one is the best model according to the following result? And why?

 Model        | Accuracy| Precision| Recall | F1 Score| ROC Score
--------------| --------| ---------|--------|---------|---------
 Random Forest| 90%     | 0.61     | 0.7    | 0.64   | 0.81      
 XGBoost      | 91%     | 0.70     | 0.67   | 0.66    | 0.80 

What I have tried:

To me, I would say Random Forest is the best model since it has a high recall ROC score.
But I am really not sure! Any answer with an explanation would be appreciated!