Principal Component Analysis (PCA) is a dimension-reduction algorithm. The idea is to use the singular value decomposition of a data matrix to obtain the directions that explain the most of the variance in the data. This boils down to finding the eigenvalue decomposition of the covariance matrix.

A toy case is shown to illustrate the fact that PCA are the directions which explain most of the variance, in order (toycase.py).

An example where PCA can be used for dimension reduction and classification is shown for the Iris dataset (irisdataset.py).

We also show how to use PCA for image compression (imagecompression.py).