pyACA.ToolPca Namespace Reference

Functions
	ToolPca (V)
	helper function: principal component analysis (pca)

Function Documentation

ToolPca()

ToolPca

(

V

)

helper function: principal component analysis (pca)

Parameters

V	features for all train observations (dimension iNumFeatures x iNumObservations)

Returns

U_pc: transformed features 'score' (dimension see V)

T: transformation matrix 'loading' (dimension iNumFeatures x iNumFeatures )

eigenvalues: 'latent' (length iNumFeatures)

Definition at line 13 of file ToolPca.py.

def ToolPca(V):
 
    # covariance
    cov_VV = np.cov(V)
 
    # svd
    [U, eigenvalues, T] = np.linalg.svd(cov_VV)
 
    # features to components
    U_pc = np.matmul(T.T, V)
 
    return U_pc, T, eigenvalues