Statistical Estimates for Generalized Splines
The authors continue the program, initiated together with \textit{S. Sun} [IEEE Trans. Automat. Control 45, 2271--2279 (2000); Math. Comput. Modeling 33, 237--253 (2001)],of producing smoothing splines as computationally feasible solutions to optimal control problems. They consider a data set of the form \[ \{(t_1,f(t_1)+\varepsilon_1),\dots,(t_N,f(t_N)+\varepsilon_N)\}, \] where \(t_1,\dots,t_N\in\mathbb R\) are the fixed sample times, \(f:\mathbb R\to\mathbb R\) is a continuous function,and \(\varepsilon_1,\dots,\varepsilon_N\) are independent, identically normally distributed random variables with mean 0 and variance \(\sigma^2\). The goal is to construct a curve using this data that is smooth and is produced from the output of the following linear, single-input, single-output control system: \[ \dot x(t)=Ax(t)+bu(t),\quad x(0)=0,\quad y(t)=cx(t)\tag{CS} \] (where \(x(t)\in\mathbb R^n\), \(u(t),y(t)\in\mathbb R\)). Under quite general conditions it is shown that \[ \lim_{N\to\infty}y^\varepsilon_N(t)=h(t) \] in probability, where \(y^\varepsilon_N\) is the response to the optimal control that minimizes \[ \tfrac1N\sum^N_{i=1}(y(t_i)-f(t_i)-\varepsilon_i)^2+\lambda\int^T_0u(t)^2\,dt \] subject to \(u\in L_2[0,T]\) and (CS), while \(h\) is the response to the optimal control that minimizes \[ \int^T_0(y(t)-f(t))^2\,dt+\lambda\int^T_0u(t)^2\,dt \] subject to the same constraints.
- Splines in Statistics
- A unified framework for spline estimators
- A unified framework for spline estimators
- scientific article; zbMATH DE number 4109290
- scientific article; zbMATH DE number 2015204
- Estimation of the probability density by spline functions
- Generalized S-Estimators
- Approximation by generalized splines
- Splines in nonparametric estimation of the density
- Spline estimation of generalised monotonic regression
- Control theoretic smoothing splines
- scientific article; zbMATH DE number 45848 (Why is no real title available?)
- Optimal control of dynamic systems: application to spline approximations
- Optimal control, statistics and path planning
- Optimal trajectory planning and smoothing splines
- Splines and linear control theory
- Splines as local smoothers
- Splines in Statistics
- Splines of class Ck on non-euclidean spaces
- The dynamic interpolation problem: on Riemannian manifolds, Lie groups, and symmetric spaces
- Approximation with the output of linear control systems
- Almost sure convergence of smoothing \(D^ m\)-splines of noisy data
- Smoothing splines for filtering strongly noised signals
- Smoothing spline via optimal control under uncertainty
- Control theoretic smoothing splines are approximate linear filters
- Optimal curve fitting and smoothing using normalized uniform B-splines: a tool for studying complex systems
- Control theoretic smoothing splines
- Optimal control, statistics and path planning
- Spline functions and stochastic filtering
This page was built for publication: Statistical Estimates for Generalized Splines
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4422730)
