Which is NOT a method to reduce overfitting?

Which is NOT a method to reduce overfitting?

• A. Cross-validation
• B. Regularization
• C. More data
• D. Increase model complexity

Answer: (D)

Overfitting happens when a model captures noise in the training data instead of the underlying pattern, so the goal is to promote generalization. Cross-validation helps by giving a more reliable estimate of how the model will perform on unseen data, guiding your choice of model complexity and hyperparameters to favor better generalization. Regularization adds a penalty to large weights, keeping the model simpler and less prone to memorizing noise. More data expands the training set, reducing variance and helping the model learn true relationships rather than random fluctuations. Increasing model complexity, however, enlarges the model’s capacity to fit the training data, including noise, which typically worsens generalization. Therefore, increasing model complexity is not a method to reduce overfitting.

Overfitting happens when a model captures noise in the training data instead of the underlying pattern, so the goal is to promote generalization. Cross-validation helps by giving a more reliable estimate of how the model will perform on unseen data, guiding your choice of model complexity and hyperparameters to favor better generalization. Regularization adds a penalty to large weights, keeping the model simpler and less prone to memorizing noise. More data expands the training set, reducing variance and helping the model learn true relationships rather than random fluctuations. Increasing model complexity, however, enlarges the model’s capacity to fit the training data, including noise, which typically worsens generalization. Therefore, increasing model complexity is not a method to reduce overfitting.