What is the primary advantage of using a regularization technique in machine learning models?

$$ It simplifies the model's parameters to a binary outcome

It helps to prevent overfitting by adding a penalty on the size of the coefficients

It allows for quicker computational times

$$ It enhances the model's ability to handle non$-$linear data