Feature Engineering Made Easy

Using the kept eigenvectors to transform new data-points

Once we decide to keep two principal components (whether we use a grid search module or the analysis of a scree plot to find the optimal number doesn't matter), we have to be able to use these components to transform incoming, out of sample data-points. To do this, let's first isolate the top two eigenvectors and store them in a new variable called top_2_eigenvectors:

# store the top two eigenvectors in a variabletop_2_eigenvectors = eig_vec_cov[:,:2].T# show the transpose so that each row is a principal component, we have two rows == two componentstop_2_eigenvectorsarray([[ 0.36158968, -0.08226889,  0.85657211,  0.35884393],
       [-0.65653988, -0.72971237,  0.1757674 ,  0.07470647]])

This array ...