in sklearn, if I want to transform the data to range(-1, 1), do you think it is better to use StandardScaler before using MinMaxScaler? to make the date more normal distributed?

  • 1
    $\begingroup$ StandardScaler() is affine linear, so there is no reason to do so. $\endgroup$
    – Michael M
    Jan 21 at 21:26
  • 2
    $\begingroup$ Both of the proposed scalers are linear transformations, and the composition of linear functions is also linear. That's addressed in the second duplicate. The first duplicate explains that neither one of these will make data resemble a normal distribution, because these scalers cannot change the shape of the data. We have many, many threads about making data more normally distributed; they can be found by using the Search feature: stats.stackexchange.com/… $\endgroup$
    – Sycorax
    Jan 21 at 21:32


Browse other questions tagged or ask your own question.