Principal Component Analysis (PCA)

Principal Component Analysis is one of the most popular dimensionality reduction techniques. Note that this algorithm is not supervised, but it is still important as a preprocessing algorithm for many other supervised techniques.

PCA computes the first \(d'\) orthogonal directions for which the data variance is maximized, where d’ is the desired dimensionality reduction.

The current PCA implementation is a wrapper for the Scikit-Learn PCA implementation.

Watch the full PCA documentation here.

Images

_images/pca.png