How to Resolve Overfitting in Machine Learning Models?
A company is developing an ML model to predict customer churn. The model performs well on the training dataset but does not accurately predict churn for new data. Which solution will resolve this issue?
Community Votes
100% of anonymous learners picked answer B. Votes are pick records left by other test-takers — they are not the verified answer.
Community Insight
This question tests your ability to diagnose overfitting and apply regularization to constrain model complexity for better generalization on unseen data.
When an ML model performs well on training data but poorly on new data, it is overfitting. Increasing the regularization parameter reduces model complexity and improves generalization.
Candidates often choose option A, thinking that more complexity will capture more patterns, but this actually worsens overfitting and further degrades performance on new data.
Community Discussion (4 comments)
Comments & Corrections
No comments yet — spotted an error or have a note? Share it below.
Expert Analysis
Diagnosing the Problem
The scenario describes a classic case of overfitting: the model has memorized the training dataset, including its noise and outliers, but fails to generalize to new, unseen data. High training accuracy combined with poor performance on new data is the hallmark of this issue.
Why Option B is Correct
Increasing the regularization parameter adds a stronger penalty for model complexity (e.g., large weights in linear models or deep trees). This forces the model to learn simpler, more general patterns rather than fitting every detail of the training data. As a result, the model's ability to generalize to new data improves significantly.
Why Other Options Are Wrong
- Option A suggests decreasing the regularization parameter, which would make the model even more complex and exacerbate overfitting.
- Option C (adding more features) increases the dimensionality of the input, which can lead to even more overfitting unless carefully managed with feature selection or dimensionality reduction.
- Option D (training for more epochs) allows the model to fit the training data even more tightly, again worsening overfitting rather than resolving it.
Community Consensus
All community voters agreed on Option B, noting that regularization is the standard technique to reduce overfitting and improve generalization.
Official Reference
Exam Strategy
When a question describes good training performance but poor new-data performance, immediately think 'overfitting.' Then look for solutions that reduce model complexity: regularization, pruning, early stopping, or less training.
Related Analysis
Practice All AIF-C01 Questions
Access 100 questions with complete answers and detailed explanations.
View Full AIF-C01 Practice Test →