MLS-C01 Question 223
Select 3You are developing a machine learning model for predicting customer churn. During the evaluation phase, you notice that your model has a high training accuracy but performs poorly on the validation dataset. Which of the following actions can help you address the issue?
- A
Collect more diverse training data to improve generalization.
- B
Increase the complexity of your model by adding more layers or features.
- C
Apply regularization techniques like L1 or L2 to reduce overfitting.
- D
Reduce the size of your training dataset to focus on higher-quality samples.
- E
Perform hyperparameter tuning to optimize model performance.
Show answer and explanation
Correct answers: A, C, E
Explanation
The scenario describes a case of overfitting, where the model performs well on the training data but poorly on the validation data. To address this, you can collect more diverse training data to improve generalization, use regularization techniques to reduce the model's complexity, and perform hyperparameter tuning to find the optimal configuration. Increasing model complexity or reducing the training dataset would likely exacerbate the problem.
- A. Correct.
Collecting more diverse training data can help reduce overfitting and improve the model's ability to generalize to unseen data, addressing the issue of high variance.
- B. Incorrect.
Increasing the complexity of your model can lead to further overfitting, worsening the issue of high variance.
- C. Correct.
Regularization techniques like L1 or L2 can penalize overly complex models, helping to reduce overfitting and improve the performance on validation data.
- D. Incorrect.
Reducing the size of the training dataset can lead to underfitting, as the model will have less data to learn from and may fail to capture important patterns.
- E. Correct.
Hyperparameter tuning can help optimize the model's configuration, potentially reducing overfitting and improving its generalization performance.