QuestionQ36

VerifiedML Model Development

An ML engineer is building a fraud-detection model by using the Amazon SageMaker XGBoost algorithm. The model categorizes transactions as fraudulent or legitimate.

During testing, the model performs very well at detecting fraud in the training dataset. However, it performs inefficiently when detecting fraud in new, unseen transactions.

What should the ML engineer do to improve fraud detection for new transactions?

Explanation

Strong training performance combined with weak performance on unseen data indicates overfitting. Decreasing max_depth reduces the complexity of the XGBoost trees, which helps the model generalize to new transactions. AWS documentation states that increasing max_depth makes the model more complex and more likely to overfit.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!