I am trying to implement logistic regression but the dataset that I have have many columns with skewed data and most of them have 0 as values. I also the skewness of data for many columns its going above 190.

But it's not only for training data, it's the same for testing data too. I tried using log method to remove skewness but because most of the value is 0 it messed up my data. I don't know how to deal with it.

I already use standarization, improved only a bit. If someone has any idea please do suggest.

  • 1
    $\begingroup$ Why shouldn’t your features be skewed? $\endgroup$
    – Dave
    Oct 11, 2021 at 7:05

1 Answer 1


If you want to fix skewness the better alternative to a simple log transform is a Power Transformation. I think Box-Cox will not work with zeros, since it accepts only positive values, but Yeo-Johnson will.

If you have a lot of zeros it might be a good idea to check for zero variance if your data is continuous and near-zero variance if your data is discrete, than delete any uninformative features.


Your Answer

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

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