The implementation of principal component analysis with the mean-squared error (MSE) minimization criterion; a.k.a. Karhunen-Loeve expantion.
The class CorPCA has a same interface as sklearn.decomposition.PCA, which has fit, transform, fit_transform inherit from sklearm.base.TransformMixin.
This package also provides a generalized PCA class corpca.PCA which takes a parameter criterion to switch the type of Gram matrix, e.g.,
from corpca import PCA
PCA(criterion='variance', n_comonents=10)
# => sklearn.decomposition.PCA(n_components=10)
PCA(criterion='mse', n_components=10)
# => corpca.CorPCA(n_components=10)
pip install .
or from PyPI
pip install corpca
MIT
Aiga SUZUKI [tochikuji@gmail.com]