0

I found this online: "StandardScaler or Z-Score Normalization is one of the feature scaling techniques, here the transformation of features is done by subtracting from the mean and dividing by standard deviation. This is often called Z-score normalization. The resulting data will have the mean as 0 and the standard deviation as 1."

To sum it up: why is it called Z-Score Normalization if it uses a Standardization technique?

What I'm thinking is that if it is called Z-Score Normalization shouldn't it use a Normalization technique rather than a Standardization one?

  • Z-Score normalization **is** standard normalization. You can view Z-score and Standard score as synonym for each other. – sagi Feb 05 '23 at 10:41
  • @VioletAster The MinMaxScaler uses the minimum and maximum value of features for scaling, not StandardScaler – Hamid Rasti Feb 05 '23 at 11:01
  • What is your definition of a "normalization technique", and how does it differ from your definition of a "standardization technique"? – Ben Reiniger Feb 05 '23 at 15:14

1 Answers1

0

According to sklearn StandardScaler documentation:

StandardScaler standardize features by removing the mean and scaling to unit variance.

The standard score of a sample x is calculated as:

z = (x - u) / s (The formula for calculating a z-score)

So, both of StandardScaler (standard normalization) and Z-Score Normalization use the same formula and they are equivalent.

Hamid Rasti
  • 813
  • 1
  • 6
  • 16
  • Thank you for your response Hamid Rasti. So in this case, the naming doesn't have any relevance? Even though is it called Z-Score Normalization, it bears no connection to Normalization? –  Feb 05 '23 at 11:04
  • Yes, the naming doesn't have any relevance. both of them are a type of normalization. and they are the same. As of @sagi commented, You can view Z-score and Standard score as synonym for each other. – Hamid Rasti Feb 05 '23 at 12:10