machine learning - AttributeError: 'StandardScaler' object has no attribute 'var_'

Question

I get this error message attempting to use the .var_ and .mean_ attribute of StandardScaler from Sci-Kit Learn. I saw in another SO post that it is no longer supported in newer versions so I downloaded an older one and it did not work either.

question from:https://stackoverflow.com/questions/65894407/attributeerror-standardscaler-object-has-no-attribute-var

Answer 1

From offical documentation the attributes var_ and mean_ are still available on the latest stable version (0.24.1 at the time of this post).

Nevertheless to access var_ and mean_ which respectively returns variance value and mean value the scaler needs to be fitted to your data. Otherwise these attributes won't be available. Also be sure that Scaler parameters with_mean and with_std are set to True. e.g. :

data = [[0, 0], [0, 0], [1, 1], [1, 1]]
scaler = StandardScaler()
scaler.mean_ ---> AttributeError: 'StandardScaler' object has no attribute 'mean_'
scaler.fit(data)
scaler.mean_ ---> array([0.5, 0.5])