Conditional least squares estimation in nonstationary nonlinear stochastic regression models

DOI10.1214/09-AOS733zbMATH Open1181.62133arXiv1001.2102MaRDI QIDQ847648FDOQ847648

Authors: Christine Jacob 

Publication date: 19 February 2010

Published in: The Annals of Statistics (Search for Journal in Brave)

Abstract: Let $Z_n$ be a real nonstationary stochastic process such that $E(Z_n|{mathcaligrF}_{n-1})stackrelmathrma.s.<infty$ and $E(Z_n^2|{mathcaligrF}_{n-1})stackrelmathrma.s.<infty$, where ${mathcaligrF}_n$ is an increasing sequence of $sigma$-algebras. Assuming that $E(Z_n|{mathcaligrF}_{n-1})=g_n(het{a}_0,u_0)=g_n^{(1)}(het{a}_0)+g_n^{(2)}(het{a}_0,u_0)$, $het{a}_{0}in{mathbbR}^p$, $p<infty$, $u_{0}in{mathbbR}^q$ and $qleqinfty$, we study the asymptotic properties of $hat{heta}_n:=argmi{n}_{heta}su{m}_{k=1}^n(Z_k-g_k(heta,hatu){)}^{2}lambd{a}_k^{-1}$, where $lambd{a}_k$ is ${mathcaligrF}_{k-1}$-measurable, $hatu=hat{u}_k$ is a sequence of estimations of $u_0$, $g_n(heta,hatu)$ is Lipschitz in $heta$ and $g_n^{(2)}(het{a}_0,hatu)-g_n^{(2)}(heta,hatu)$ is asymptotically negligible relative to $g_n^{(1)}(het{a}_0)-g_n^{(1)}(heta)$. We first generalize to this nonlinear stochastic model the necessary and sufficient condition obtained for the strong consistency of $hat{heta}_n$ in the linear model. For that, we prove a strong law of large numbers for a class of submartingales. Again using this strong law, we derive the general conditions leading to the asymptotic distribution of $hat{heta}_n$. We illustrate the theoretical results with examples of branching processes, and extension to quasi-likelihood estimators is also considered.

Full work available at URL: https://arxiv.org/abs/1001.2102

Recommendations

• Asymptotic properties of nonlinear least squares estimates in stochastic regression models
• On conditional variance estimation in nonparametric regression
• Estimation in semi-parametric regression with non-stationary regressors
• Robust Inference in Conditionally Linear Nonlinear Regression Models
• Nonlinear regression of stable random variables
• Estimation and inference in unstable nonlinear least squares models
• scientific article; zbMATH DE number 799018
• scientific article; zbMATH DE number 1843490

zbMATH Keywords

heteroscedasticityconsistencytime seriesasymptotic distributionbranching processconditional least squares estimatorsubmartingalenonstationary processmartingale differencepolymerase chain reactionquasi-likelihood estimatorstochastic nonlinear regression

Mathematics Subject Classification ID

Asymptotic properties of parametric estimators (62F12) General nonlinear regression (62J02) Markov processes: estimation; hidden Markov models (62M05) Time series, auto-correlation, regression, etc. in statistics (GARCH) (62M10) Asymptotic distribution theory in statistics (62E20) Non-Markovian processes: estimation (62M09) Strong limit theorems (60F15) Martingales and classical analysis (60G46)

Cites Work

• Autoregressive Conditional Heteroscedasticity with Estimates of the Variance of United Kingdom Inflation
• Title not available (Why is that?)
• Quasi-Likelihood and Optimal Estimation, Correspondent Paper
• Asymptotic Properties of Non-Linear Least Squares Estimators
• Central limit theorems for local martingales
• Asymptotic properties of general autoregressive models and strong consistency of least-squares estimates of their parameters
• Title not available (Why is that?)
• Title not available (Why is that?)
• Least squares estimates in stochastic regression models with applications to identification and control of dynamic systems
• On population-size-dependent branching processes
• Title not available (Why is that?)
• Inference in Arch and Garch Models with Heavy-Tailed Errors
• Asymptotic theory of nonlinear least squares estimation
• Quasi-likelihood estimation for semimartingales
• Title not available (Why is that?)
• Asymptotic properties of nonlinear least squares estimates in stochastic regression models
• Branching Processes
• Estimation in a class of nonlinear heteroscedastic time series models
• Strong consistency of least squares estimates in dynamic models
• Strong consistency in nonlinear stochastic regression models.
• Asymptotic properties of least-squares estimates in stochastic regression models
• Strong consistency of least squares estimates in multiple regression II
• On least squares estimation for stable nonlinear AR processes
• Estimation of the parameters of a branching process from migrating binomial observations
• Random sums and branching stochastic processes
• Estimation of the offspring mean in a supercritical or near-critical size-dependent branching process
• Modelling the PCR amplification process by a size-dependent branching process and estimation of the efficiency
• Estimation of the mean and the initial probabilities of a branching process
• Identification of multitype branching processes
• A NOTE ON MAXIMUM LIKELIHOOD ESTIMATION FOR SIMPLE BRANCHING PROCESSES

Cited In (11)

• Asymptotic properties of nonlinear least squares estimates in stochastic regression models
• Weighted nonlinear regression with nonstationary time series
• Title not available (Why is that?)
• A generalized quasi-likelihood estimator for nonstationary stochastic processes -- asymptotic properties and examples.
• Quasi- and pseudo-maximum likelihood estimators for discretely observed continuous-time Markov branching processes
• Least squares estimation for nonlinear regression models with heteroscedasticity
• Estimation of the infection parameter of an epidemic modeled by a branching process
• Regression analysis of stochastic fatigue crack growth model in a martingale difference framework
• Stochastic methodology for the study of an epidemic decay phase, based on a branching model
• Conditional least squares estimation for the SINAR(1, 1) process
• Conditional moment estimation of nonlinear equation systems

This page was built for publication: Conditional least squares estimation in nonstationary nonlinear stochastic regression models

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q847648)