ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 1312.3120
45
2

Marked empirical processes for non-stationary time series

11 December 2013
N. Chan
Rongmao Zhang
ArXiv (abs)PDFHTML
Abstract

Consider a first-order autoregressive process Xi=βXi−1+εi,X_i=\beta X_{i-1}+\varepsilon_i,Xi​=βXi−1​+εi​, where εi=G(ηi,ηi−1,…)\varepsilon_i=G(\eta_i,\eta_{i-1},\ldots)εi​=G(ηi​,ηi−1​,…) and ηi,i∈Z\eta_i,i\in\mathbb{Z}ηi​,i∈Z are i.i.d. random variables. Motivated by two important issues for the inference of this model, namely, the quantile inference for H0:β=1H_0: \beta=1H0​:β=1, and the goodness-of-fit for the unit root model, the notion of the marked empirical process αn(x)=1n∑i=1ng(Xi/an)I(εi≤x),x∈R\alpha_n(x)=\frac{1}{n}\sum_{i=1}^ng(X_i/a_n)I(\varepsilon_i\leq x),x\in\mathbb{R}αn​(x)=n1​∑i=1n​g(Xi​/an​)I(εi​≤x),x∈R is investigated in this paper. Herein, g(⋅)g(\cdot)g(⋅) is a continuous function on R\mathbb{R}R and {an}\{a_n\}{an​} is a sequence of self-normalizing constants. As the innovation {εi}\{\varepsilon_i\}{εi​} is usually not observable, the residual marked empirical process α^n(x)=1n∑i=1ng(Xi/an)I(ε^i\leqx),x∈R,\hat {\alpha}_n(x)=\frac{1}{n}\sum_{i=1}^ng(X_i/a_n)I(\hat{\varepsilon}_i\l eq x),x\in\mathbb{R},α^n​(x)=n1​∑i=1n​g(Xi​/an​)I(ε^i​\leqx),x∈R, is considered instead, where ε^i=Xi−β^Xi−1\hat{\varepsilon}_i=X_i-\hat{\beta}X_{i-1}ε^i​=Xi​−β^​Xi−1​ and β^\hat{\beta}β^​ is a consistent estimate of β.\beta.β. In particular, via the martingale decomposition of stationary process and the stochastic integral result of Jakubowski (Ann. Probab. 24 (1996) 2141-2153), the limit distributions of αn(x)\alpha_n(x)αn​(x) and α^n(x)\hat{\alpha}_n(x)α^n​(x) are established when {εi}\{\varepsilon_i\}{εi​} is a short-memory process. Furthermore, by virtue of the results of Wu (Bernoulli 95 (2003) 809-831) and Ho and Hsing (Ann. Statist. 24 (1996) 992-1024) of empirical process and the integral result of Mikosch and Norvai\v{s}a (Bernoulli 6 (2000) 401-434) and Young (Acta Math. 67 (1936) 251-282), the limit distributions of αn(x)\alpha_n(x)αn​(x) and α^n(x)\hat{\alpha}_n(x)α^n​(x) are also derived when {εi}\{\varepsilon_i\}{εi​} is a long-memory process.

View on arXiv
Comments on this paper