Databricks Machine Learning Associate Question 618
Single answerYou are training a machine learning model to predict customer churn for a subscription-based service. During the model evaluation, you notice that your model has high training accuracy but performs poorly on the test set. Which of the following scenarios best explains this issue and how it relates to the bias-variance tradeoff?
- A
The model is underfitting due to high bias, and increasing model complexity will improve its performance.
- B
The model is overfitting due to high variance, and regularization techniques or reducing model complexity may help.
- C
The model is underfitting due to high variance, and using a larger training dataset will improve its performance.
- D
The model is overfitting due to high bias, and increasing the number of features will improve its performance.
Show answer and explanation
Correct answer: B
Explanation
The question addresses the bias-variance tradeoff and its impact on model performance. High training accuracy and poor test accuracy suggest that the model is overfitting the training data, which is a sign of high variance. Overfitting can be mitigated by using techniques such as regularization, simplifying the model, or other methods to improve generalization. Understanding how to recognize and address overfitting is critical for effectively managing the bias-variance tradeoff.
- A. Incorrect.
This is incorrect. High training accuracy suggests the model is not underfitting. The issue is more likely related to overfitting due to high variance.
- B. Correct.
This is correct. High training accuracy combined with poor test performance indicates the model is overfitting the training data, which is a symptom of high variance. Techniques such as regularization or reducing model complexity can help mitigate this issue by improving generalization.
- C. Incorrect.
This is incorrect. High variance refers to overfitting rather than underfitting. Increasing the training dataset might help in some cases, but it does not directly address the high variance issue.
- D. Incorrect.
This is incorrect. Overfitting is caused by high variance, not high bias. Additionally, increasing the number of features often increases model complexity, which could exacerbate overfitting rather than resolve it.