Search code examples
pythonneural-networktraining-datastandard-deviation

Scale training set by inverse standard deviation using python

I am using python and have a training set of data that i need to 'subtract the mean and scale by inverse standard deviation'. Subtracting the mean would just be subtracting the mean from each value in each column i assume, but i have no idea what i am meant to do when it says to 'scale by inverse standard deviation'.

I have googled it but nothing has come up in relation to python or neural networks so i'm not sure how to continue.

Thanks

EDIT: Would this be correct?

scaled_train =  (train - train_mean) / train_std_deviation
Solution

  • In the future these questions are better for CrossValidated.

    Let your dataset be x then 

    import numpy as np
x = np.array(x)
x -= np.mean(x)
x /= x.std()

    This is called Standardization

    This can be achieved with sklearn as per the docs for 

    >>> from sklearn import preprocessing
>>> import numpy as np
>>> X_train = np.array([[ 1., -1.,  2.],
...                     [ 2.,  0.,  0.],
...                     [ 0.,  1., -1.]])
>>> X_scaled = preprocessing.scale(X_train)

    References