Preprocessing

Before the analysis, functional data need to be preprocessed. The package provides several classes to preprocess functional data, including smoothing and dimension reduction.

Smoothing

Observations of functional data are often noisy. Smoothing methods are used to remove the noise and extract the underlying patterns. The package provides two classes to smooth functional data: LocalPolynomial and PSplines.

FDApy.preprocessing.LocalPolynomial([...])	Local Polynomial regression.
FDApy.preprocessing.PSplines([n_segments, ...])	P-Splines smoothing.

Dimension reduction

Due to the infinite-dimensional nature of functional data, dimension reduction techniques are important tools in functional data analysis. They are used to extract the most relevant information. The package provides three classes to reduce the dimension of functional data: UFPCA, MFPCA, and FCPTPA.

FDApy.preprocessing.UFPCA([method, ...])	Univariate functional principal components analysis.
FDApy.preprocessing.MFPCA([n_components, ...])	Multivariate functional principal components analysis.
FDApy.preprocessing.FCPTPA([n_components, ...])	Functional canonical polyadic-tensor power algorithm.