Mahalanobis distance, covariance estimation, and Principal Component Analysis (PCA) —
mirrors scipy.spatial.distance.mahalanobis and sklearn.decomposition.PCA.
mahalanobis(u, v, VI) computes distance using inverse covariance VI.
Use covMatrix + invertMatrix to estimate VI from data.
new PCA({ n_components }) reduces dimensionality. fit returns the result with
explained_variance_ratio, components, and transform.
Reduce 3-dimensional data to 2 components and inspect how much variance each component captures.