Feature Engineering Made Easy

Creating the covariance matrix of the dataset

To calculate the covariance matrix of iris, we will first calculate the feature-wise mean vector (for use in the future) and then calculate our covariance matrix using NumPy.

The covariance matrix is a d x d matrix (square matrix with the same number of features as the number of rows and columns) that represents feature interactions between each feature. It is quite similar to a correlation matrix:

# Calculate a PCA manually# import numpyimport numpy as np# calculate the mean vectormean_vector = iris_X.mean(axis=0)print mean_vector[ 5.84333333  3.054       3.75866667  1.19866667]
# calculate the covariance matrixcov_mat = np.cov((iris_X-mean_vector).T)print cov_mat.shape(4, 4)

The variable cov_mat stores ...