MLS-C01 Question 85
Select 3You are developing a machine learning model to predict customer churn using Amazon SageMaker. During testing, you observe that your model performs well on the training dataset but poorly on the validation dataset, indicating overfitting. Which of the following actions can help mitigate overfitting in this scenario?
- A
Apply L1 or L2 regularization to the model.
- B
Increase the size of your training dataset.
- C
Use a deeper neural network architecture to improve model complexity.
- D
Implement dropout during training.
- E
Reduce the learning rate of your optimizer.
Show answer and explanation
Correct answers: A, B, D
Explanation
Overfitting occurs when a model learns patterns specific to the training data but fails to generalize to unseen data. Techniques such as L1/L2 regularization, increasing the training dataset size, and implementing dropout are effective strategies to mitigate overfitting. These methods either constrain the model's complexity or provide more diverse data to improve generalization. Using a deeper architecture or reducing the learning rate does not directly address the issue of overfitting.
- A. Correct.
Applying L1 or L2 regularization can penalize large weights in the model, reducing overfitting and encouraging simpler models.
- B. Correct.
Increasing the size of your training dataset can help the model generalize better by providing more diverse data points to learn from.
- C. Incorrect.
Using a deeper neural network increases the complexity of the model, which can worsen overfitting if the model is already too complex.
- D. Correct.
Dropout randomly deactivates neurons during training, which prevents the model from relying too heavily on specific neurons and reduces overfitting.
- E. Incorrect.
Reducing the learning rate can help with convergence issues but does not directly address overfitting.