1
$\begingroup$

I would like to know the use of correaltion map in machine leraning. For example, if there are 2 features with high correaltion, should either of the features be removed before appying the algorithm or it depends on every data set. Any explanation would be highly helpful. Thanks in advance.

$\endgroup$
1
$\begingroup$

It depends. A high correlation between two features suggests that they represents almost the same information. For some problems like clustering, it is always useful to remove redundant features while some algorithm like Gradient Boosting in xgboost is not affected at all by such features. So, it all depends on what you want to do with your data set.

As per my opinion, if your dataset has too many features, then I would suggest to check the correlation between those features and apply PCA to reduce the dimensionality of your dataset especially if you are doing tasks like clustering or regression.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.