pca#

diffsptk.PCA#: alias of PrincipalComponentAnalysis

class diffsptk.PrincipalComponentAnalysis(n_comp, cov_type='sample')[source]#

See this page for details.

Parameters:

n_compint >= 1 [scalar]: Number of principal components, \(N\).
cov_type[‘sample’, ‘unbiased’, ‘correlation’]: Type of covariance.

forward(x)[source]#

Perform PCA.

Parameters:

xTensor [shape=(…, M+1)]: Input vectors.

Returns:

eTensor [shape=(N,)]: Eigenvalues ordered in ascending order.
vTensor [shape=(M+1, N)]: Eigenvectors.
mTensor [shape=(M+1,)]: Mean vector.

Examples

>>> x = diffsptk.nrand(10, 3)
>>> pca = diffsptk.PCA(3)
>>> e, _, _ = pca(x)
>>> e
tensor([0.6240, 1.0342, 1.7350])