Developer Guide for Intel® Data Analytics Acceleration Library 2018 Update 3

Z-score

Z-score normalization is an algorithm to normalize the observations by each feature (column).

Problem Statement

Given a set X of n feature vectors x 1 = (x 11 , … , x 1p ), ... , x n = (x n1 , … , x np ) of dimension p, the problem is to compute the matrix Y = (y i j ) n x p where the j-th column (Y ) j = (y i j ) i= 1, ..., n is obtained as a result of normalizing the column (X ) j = (x i j ) i= 1, ..., n of the original matrix as:

where m j is the mean and σ j is the standard deviation of column (X ) j .