What role does dimensionality reduction play in data preprocessing for machine learning?

$$ It enhances the model's ability to handle large datasets.

It increases the diversity of the training data.

It reduces the number of input variables to improve model performance and prevent overfitting.

$$ It simplifies the computational requirements for model training.