MLA-C01 Question 190
Select 2You are building a classification model using Amazon SageMaker's XGBoost algorithm. During the initial training, you observe that the model is underfitting the training data. Which of the following hyperparameter adjustments could help improve the model's performance?
- A
Increase the number of trees in the model
- B
Decrease the learning rate
- C
Reduce the maximum depth of the trees
- D
Increase the maximum depth of the trees
- E
Reduce the number of features considered for each split
Show answer and explanation
Correct answers: A, D
Explanation
Underfitting occurs when a model is too simple to capture the underlying patterns in the data. To address this, you can increase the model complexity. In the case of XGBoost, increasing the number of trees and the depth of the trees are effective strategies to enhance the model's ability to learn from the data. Care must be taken to monitor for overfitting when making these adjustments.
- A. Correct.
Increasing the number of trees allows the model to better learn from the data, potentially reducing underfitting. However, this should be done with care to avoid overfitting.
- B. Incorrect.
Decreasing the learning rate slows down the training process and can help in fine-tuning, but it is unlikely to directly address underfitting as it does not significantly increase model complexity.
- C. Incorrect.
Reducing the maximum depth of the trees makes the model simpler, which could exacerbate underfitting rather than improving it.
- D. Correct.
Increasing the maximum depth of the trees allows the model to capture more complex patterns in the data, helping to reduce underfitting.
- E. Incorrect.
Reducing the number of features considered for each split reduces model complexity, which is counterproductive when addressing underfitting.