Can anyone explain me StandardScaler?
Intro
I assume that you have a matrix X
where each row/line is a sample/observation and each column is a variable/feature (this is the expected input for any sklearn
ML function by the way -- X.shape
should be [number_of_samples, number_of_features]
).
Core of method
The main idea is to normalize/standardize i.e. μ = 0
and σ = 1
your features/variables/columns of X
, individually, before applying any machine learning model.
StandardScaler()
will normalize the features i.e. eachcolumn of X, INDIVIDUALLY, so that each column/feature/variable will haveμ = 0
andσ = 1
.
P.S: I find the most upvoted answer on this page, wrong.I am quoting "each value in the dataset will have the sample mean value subtracted" -- This is neither true nor correct.
See also: How and why to Standardize your data: A python tutorial
Example with code
from sklearn.preprocessing import StandardScalerimport numpy as np# 4 samples/observations and 2 variables/featuresdata = np.array([[0, 0], [1, 0], [0, 1], [1, 1]])scaler = StandardScaler()scaled_data = scaler.fit_transform(data)print(data)[[0, 0], [1, 0], [0, 1], [1, 1]])print(scaled_data)[[-1. -1.] [ 1. -1.] [-1. 1.] [ 1. 1.]]
Verify that the mean of each feature (column) is 0:
scaled_data.mean(axis = 0)array([0., 0.])
Verify that the std of each feature (column) is 1:
scaled_data.std(axis = 0)array([1., 1.])
Appendix: The maths
UPDATE 08/2020: Concerning the input parameters with_mean
and with_std
to False
/True
, I have provided an answer here: StandardScaler difference between “with_std=False or True” and “with_mean=False or True”
The idea behind StandardScaler
is that it will transform your data such that its distribution will have a mean value 0 and standard deviation of 1.
In case of multivariate data, this is done feature-wise (in other words independently for each column of the data).
Given the distribution of the data, each value in the dataset will have the mean value subtracted, and then divided by the standard deviation of the whole dataset (or feature in the multivariate case).
How to calculate it:
You can read more here: