Why is it sometimes recommended to remove features that are highly correlated with each other?
a) To reduce the risk of overfitting and improve model generalization
b) To increase the number of features available for analysis
c) To make the dataset more complex and comprehensive
d) To improve the interpretability of the model