API Reference

Complete reference for all fdars modules.

from fdars import Fdata  # main entry point

Modules

Module	Description
Fdata	Functional data container — the main entry point (1D curves, 2D surfaces, metadata, methods)
fdata	Low-level functional data operations: mean, centering, derivatives, norms, median, normalization
depth	Depth measures for functional data (Fraiman-Muniz, modal, band, projection, spatial)
metric	Distance metrics and matrices (Lp, Hausdorff, DTW, Soft-DTW, Fourier, horizontal shift)
basis	Basis representations: B-spline, Fourier, P-spline fitting, automatic selection
smoothing	Nonparametric smoothing: Nadaraya-Watson, local polynomial, k-NN, bandwidth selection
clustering	Clustering: k-means, fuzzy C-means, GMM, silhouette, Calinski-Harabasz
regression	Regression: FPCA, FPLS, scalar-on-function, function-on-scalar, ANOVA, logistic
alignment	Elastic alignment and shape analysis: SRSF, Karcher mean, warping, elastic FPCA
outliers	Outlier detection: LRT bootstrap, outliergram, magnitude-shape
seasonal	Seasonal analysis: period detection (SAZED, autoperiod), STL decomposition, peaks
spm	Statistical Process Monitoring: Phase I/II control charts, Hotelling T-squared
classification	Classification: LDA, QDA, k-NN, kernel, cross-validation
tolerance	Tolerance and confidence bands: FPCA-based, conformal, Degras SCB, equivalence test
conformal	Conformal prediction: regression intervals, nonparametric, classification sets
simulation	Simulation: Karhunen-Loeve expansion, Gaussian processes, covariance matrices
explain	Explainability: permutation importance, PDP, SHAP values, significant regions

Conventions

• Array inputs: All array parameters accept numpy.ndarray. Data matrices have shape (n_obs, n_points).
• Return types: Functions return numpy.ndarray, float, dict, or list depending on the output.
• Optional parameters: Shown with =default in signatures. Pass None for auto-selection where noted.
• Errors: Invalid inputs raise ValueError with a descriptive message.