3
$\begingroup$

Can someone explain to me like I'm five on why multicollinearity does not affect neural networks?

I've done some research and neural networks are basically linear functions being stacked with activation functions in between, now if the original input variables are highly correlated, doesn't that mean multicollinearity happens?

Improve this question
$\endgroup$

1 Answer 1

Reset to default
5
$\begingroup$

Multicollinearity in linear or logistic regression does not impact the model performance it only impacts how model coefficient are interpreted. Multicollinearity leads to a problem where small change in one variable can lead to drastic changes in coefficients.

A neural network is black box model in nature so if its performs to a given expectation (good accuracy) we never know the impact of multicollinearity. Thats why i think most blogs say that neural network are not affected by multicollinearity

Improve this answer
$\endgroup$
1
  • 1
    $\begingroup$ So basically all we care about is prediction accuracy and not the interpretability, hence it is disregarded $\endgroup$
    – Chukwudi Ogbonna
    Commented Mar 4, 2022 at 16:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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