Abstract:We propose several methods that address the problem of fitting a C1\documentclass[12pt]{minimal}\usepackage{amsmath}\usepackage{wasysym}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{mathrsfs}\usepackage{upgreek}\setlength{\oddsidemargin}{-69pt}\begin{document}$$C^1$$\end{document} curve γ\documentclass[12pt]{minimal}\usepackage{amsmath}\usepackage{wasysym}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{mathrsfs}\usepackage{upgreek}\setlength{\oddsidemargin}{-69pt}\begin{document}$$\gamma $$\end{document} to time-labeled data points on a manifold. The methods have a parameter, λ\documentclass[12pt]{minimal}\usepackage{amsmath}\usepackage{wasysym}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{mathrsfs}\usepackage{upgreek}\setlength{\oddsidemargin}{-69pt}\begin{document}$$\lambda $$\end{document}, to adjust the relative importance of the two goals that the curve should meet: being “straight enough” while fitting the data “closely enough.” The methods are designed for ease of use: they only require to compute Riemannian exponentials and logarithms, they represent the curve by means of a number of tangent vectors that grows linearly with the number of data points, and, once the representation is computed, evaluating γ(t)\documentclass[12pt]{minimal}\usepackage{amsmath}\usepackage{wasysym}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{mathrsfs}\usepackage{upgreek}\setlength{\oddsidemargin}{-69pt}\begin{document}$$\gamma (t)$$\end{document} at any t requires a small number of exponentials and logarithms (independent of the number of data points). Among the proposed methods, the blended cubic spline technique combines the additional properties of interpolating the data when λ→∞\documentclass[12pt]{minimal}\usepackage{amsmath}\usepackage{wasysym}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{mathrsfs}\usepackage{upgreek}\setlength{\oddsidemargin}{-69pt}\begin{document}$$\lambda \rightarrow \infty $$\end{document} and reducing to the well-known cubic smoothing spline when the manifold is Euclidean. The methods are illustrated on synthetic and real data.

Very smooth curve estimation via semiparametric penalty

Smooth Exact Penalty Functions: A General Approach

Dose-Response Curve Estimation: A Semiparametric Mixture Approach

An Efficient Penalized Estimation Approach for Semiparametric Linear Transformation Models with Interval‐censored Data

Estimating Functions and Derivatives Via Adaptive Penalized Splines.

Fast Bivariate Penalized Splines: the Sandwich Smoother

Multiple smoothing parameters selection in additive regression quantiles

Minimal penalty for Goldenshluger-Lepski method

Efficient Estimation in Single Index Models through Smoothing splines

Smooth exact penalty functions II: a reduction to standard exact penalty functions

Stable and Efficient Multiple Smoothing Parameter Estimation for Generalized Additive Models

A smoothed semiparametric likelihood for estimation of nonparametric finite mixture models with a copula-based dependence structure

Fitting Smoothing Splines to Time-Indexed, Noisy Points on Nonlinear Manifolds

Data Fitting on Manifolds with Composite Bézier-Like Curves and Blended Cubic Splines

Asymptotic Properties of Penalized Spline Estimators in Concave Extended Linear Models: Rates of Convergence

A Monte Carlo Method for Variance Estimation for Estimators Based on Induced Smoothing.

Adaptive shape-preserving curve fitting on surfaces

Smoothing Spline ANOVA Models

Smoothing Sparse and Unevenly Sampled Curves using Semiparametric Mixed Models: An Application to Online Auctions

Semi-parametric estimation of shifts

Gradient-type penalty method with inertial effects for solving constrained convex optimization problems with smooth data