Threshold Autoregressive Process Research Articles

A new integer-valued threshold autoregressive process based on modified negative binomial operator driven by explanatory variables

A new integer-valued threshold autoregressive process based on modified negative binomial operator driven by explanatory variables

Almost Sure Central Limit Theorem for Error Variance Estimator in Pth-Order Nonlinear Autoregressive Processes

In this paper, under some suitable assumptions, using the Taylor expansion, Borel–Cantelli lemma and the almost sure central limit theorem for independent random variables, the almost sure central limit theorem for error variance estimator in the pth-order nonlinear autoregressive processes with independent and identical distributed errors was established. Four examples, first-order autoregressive processes, self-exciting threshold autoregressive processes, threshold-exponential AR progresses and multilayer perceptrons progress, are given to verify the results.

A Time-Varying Mixture Integer-Valued Threshold Autoregressive Process Driven by Explanatory Variables.

In this paper, a time-varying first-order mixture integer-valued threshold autoregressive process driven by explanatory variables is introduced. The basic probabilistic and statistical properties of this model are studied in depth. We proceed to derive estimators using the conditional least squares (CLS) and conditional maximum likelihood (CML) methods, while also establishing the asymptotic properties of the CLS estimator. Furthermore, we employed the CLS and CML score functions to infer the threshold parameter. Additionally, three test statistics to detect the existence of the piecewise structure and explanatory variables were utilized. To support our findings, we conducted simulation studies and applied our model to two applications concerning the daily stock trading volumes of VOW.

Drift Estimation of the Threshold Ornstein-Uhlenbeck Process From Continuous and Discrete Observations

We refer by threshold Ornstein-Uhlenbeck to a continuous-time threshold autoregressive process. It follows the Ornstein-Uhlenbeck dynamics when above or below a fixed level, yet at this level (threshold) its coefficients can be discontinuous. We discuss (quasi)-maximum likelihood estimation of the drift parameters, both assuming continuous and discrete time observations. In the ergodic case, we derive consistency and speed of convergence of these estimators in long time and high frequency. Based on these results, we develop a test for the presence of a threshold in the dynamics. Finally, we apply these statistical tools to short-term US interest rates modeling.

A new First-Order mixture integer-valued threshold autoregressive process based on binomial thinning and negative binomial thinning

In this paper, we introduce a new first-order mixture integer-valued threshold autoregressive process, based on the binomial and negative binomial thinning operators. Basic probabilistic and statistical properties of this model are discussed. Conditional least squares (CLS) and conditional maximum likelihood (CML) estimators are derived and the asymptotic properties of the estimators are established. The inference for the threshold parameter is obtained based on the CLS and CML score functions. Moreover, the Wald test is applied to detect the existence of the piecewise structure. Simulation studies are considered, along with an application: the number of criminal mischief incidents in the Pittsburgh dataset

On MCMC sampling in random coefficients self-exciting integer-valued threshold autoregressive processes

In this study, Bayesian estimation is performed for a class of random coefficient self-exciting integer-valued threshold autoregressive processes with explanatory variables. A new model with a linear structure is obtained through model reconstruction, which makes Markov Chain Monte Carlo method easy to perform. By introducing the latent variables series, a complete data likelihood is obtained. Based on this likelihood, the full conditional distributions are easily obtained for all the parameters and latent variables. By maximizing the posterior probability function, the threshold parameter is accurately estimated. Finally, some numerical results of the estimates and a real data example of crime counts in Ballina, New South Wales, Australia are presented.

A nonparametric Bayesian analysis for meningococcal disease counts based on integer-valued threshold time series models

To better describe the characteristics of time series of counts showing over-dispersion, asymmetry and structural change features, this paper considers a class of random coefficient integer-valued threshold autoregressive processes that properly capture flexible asymmetric and nonlinear responses without assuming the distributions for the errors. By the Taylor expansion, a normally distributed working likelihood is obtained and the Bayesian empirical likelihood inference is conducted for the model parameters. Through the normalized approximation of the nonparametric likelihood, the originally complicated Bayesian calculation has been greatly simplified. Finally, the proposed method is verified via some simulations and an empirical analysis of the meningococcal disease data set in Germany.

An alternative sequential method for the state estimation of a partially observed SETAR(1) process

We discuss a novel sequential test based on the quadratic variation of the observations to decide which regime governs the dynamics of the non-observed process in a filtering problem with small observation noise. The non-observed state process is a self-exciting threshold autoregressive process of order one (SETAR(1)) with two regimes. The observation function is not one-to-one. The proposed procedure performs well and may be competitive in some applications.

Penalized estimation of threshold auto-regressive models with many components and thresholds.

Thanks to their simplicity and interpretable structure, autoregressive processes are widely used to model time series data. However, many real time series data sets exhibit non-linear patterns, requiring nonlinear modeling. The threshold Auto-Regressive (TAR) process provides a family of non-linear auto-regressive time series models in which the process dynamics are specific step functions of a thresholding variable. While estimation and inference for low-dimensional TAR models have been investigated, high-dimensional TAR models have received less attention. In this article, we develop a new framework for estimating high-dimensional TAR models, and propose two different sparsity-inducing penalties. The first penalty corresponds to a natural extension of classical TAR model to high-dimensional settings, where the same threshold is enforced for all model parameters. Our second penalty develops a more flexible TAR model, where different thresholds are allowed for different auto-regressive coefficients. We show that both penalized estimation strategies can be utilized in a three-step procedure that consistently learns both the thresholds and the corresponding auto-regressive coefficients. However, our theoretical and empirical investigations show that the direct extension of the TAR model is not appropriate for high-dimensional settings and is better suited for moderate dimensions. In contrast, the more flexible extension of the TAR model leads to consistent estimation and superior empirical performance in high dimensions.

Random coefficients integer-valued threshold autoregressive processes driven by logistic regression

In this article, we introduce a new random coefficients self-exciting threshold integer-valued autoregressive process. The autoregressive coefficients are driven by a logistic regression structure, so that the explanatory variables can be included. Basic probabilistic and statistical properties of this model are discussed. Conditional least squares and conditional maximum likelihood estimators, as well as the asymptotic properties of the estimators, are discussed. The nonlinearity test of the model and existence test of explanatory variables are also addressed. As an illustration, we evaluate our estimates through a simulation study. Finally, we apply our method to the data sets of sexual offences in Ballina, New South Wales (NSW), Australia, with two covariates of temperature and drug offences. The result reveals that the proposed model fits the data sets well.

Bayesian Analysis of Multiplicative Seasonal Threshold Autoregressive Processes

Seasonal fluctuations are often found in many time series. In addition, non-linearity and the relationship with other time series are prominent behaviors of several, of such series. In this paper, we consider the modeling of multiplicative seasonal threshold autoregressive processes with exogenous input (TSARX), which explicitly and simultaneously incorporate multiplicative seasonality and threshold nonlinearity. Seasonality is modeled to be stochastic and regime dependent. The proposed model is a special case of a threshold autoregressive process with exogenous input (TARX). We develop a procedure based on Bayesian methods to identify the model, estimate parameters, validate the model and calculate forecasts. In the identification stage of the model, we present a statistical test of regime dependent multiplicative seasonality. The proposed methodology is illustrated with a simulated example and applied to economic empirical data.

A perturbation analysis of Markov chains models with time-varying parameters

We study some regularity properties in locally stationary Markov models which are fundamental for controlling the bias of nonparametric kernel estimators. In particular, we provide an alternative to the standard notion of derivative process developed in the literature and that can be used for studying a wide class of Markov processes. To this end, for some families of $V$-geometrically ergodic Markov kernels indexed by a real parameter $u$, we give conditions under which the invariant probability distribution is differentiable with respect to $u$, in the sense of signed measures. Our results also complete the existing literature for the perturbation analysis of Markov chains, in particular when exponential moments are not finite. Our conditions are checked on several original examples of locally stationary processes such as integer-valued autoregressive processes, categorical time series or threshold autoregressive processes.

An empirical study on the parsimony and descriptive power of TARMA models

In linear time series analysis, the incorporation of the moving-average term in autoregressive models yields parsimony while retaining flexibility; in particular, the first order autoregressive moving-average model, ARMA(1,1) is notable since it retains a good approximating capability with just two parameters. In the same spirit, we assess empirically whether a similar result holds for threshold processes. First, we show that the first order threshold autoregressive moving-average process, TARMA(1,1) exhibits complex, high-dimensional, behaviour with parsimony, by comparing it with threshold autoregressive processes, TAR(p), with possibly large autoregressive order p. Second, we study the descriptive power of the TARMA(1,1) model with respect to the class of autoregressive models, seen as universal approximators: in several situations, the TARMA(1,1) model outperforms AR(p) models even when p is large. Lastly, we analyze two real world data sets: the sunspot number and the male US unemployment rate time series. In both cases, we show that TARMA models provide a better fit with respect to the best TAR models proposed in literature.

On the Stationary Marginal Distributions of Subclasses of Multivariate Setar Processes of Order One

To introduce more flexibility in process‐parameters through a regime‐switching behavior, the classical autoregressive (AR) processes have been extended to self‐exciting threshold autoregressive (SETAR) processes. However, the stationary marginal distributions of SETAR processes are usually difficult to obtain in explicit forms and, therefore, they lack appropriate characterizations. The stationary marginal distribution of a multivariate (‐dimensional) SETAR process of order one with multivariate normal innovations is shown to belong to the unified skew‐normal () family and characterized under a fairly broad condition. This article also characterizes the stationary marginal distributions of a subclass of the processes with symmetric multivariate stable innovations. To characterize the stationary marginal distributions of these processes, the authors show that they belong to specific skew‐distribution families, and for a given skew‐distribution from the corresponding family, an process, with stationary marginal distribution identical to the given skew‐distribution, can be associated. Furthermore, this article illustrates a diagnostic of an model using the corresponding stationary marginal density.

Martingale decomposition and approximations for nonlinearly dependent processes

This paper proposes a new martingale (MG) decomposition (Gordin, 1969; Hall and Heyde, 1980) for a dependent time series under a predictive dependence measure based on Wu (2005). The decomposition produces a generalized version of the Beveridge–Nelson (BN) lemma (Phillips and Solo, 1992) that accommodates many nonlinear time series, such as GARCH models and threshold autoregressive processes, thereby extending the empirical ambit of the original lemma designed for the linear process. Under this extended framework, MG approximations can be constructed for weighted sums of the nonlinear dependent processes and these approximations lead directly to a new central limit theorem whose range of application includes many practical time series models.

Martingale Decomposition and Approximations for Nonlinearly Dependent Processes

This paper proposes a new martingale (MG) decomposition (Gordin, 1969; Hall and Heyde, 1980) for a dependent time series under a predictive dependence measure based on Wu (2005). The decomposition produces a generalized version of the Beveridge-Nelson (BN) lemma (Phillips and Solo, 1992) that accommodates many nonlinear time series, such as GARCH models and threshold autoregressive processes, thereby extending the empirical ambit of the original lemma designed for the linear process. Under this extended framework, MG approximations can be constructed for weighted sums of the nonlinear dependent processes and these approximations lead directly to a new central limit theorem whose range of application includes many practical time series models.

Estimation of parameters in the self-exciting threshold autoregressive processes for nonlinear time series of counts

To better describe the characteristics of time series of counts such as over-dispersion, asymmetry and structural change, this paper considers a class of integer-valued self-exciting threshold autoregressive processes that properly capture flexible asymmetric and nonlinear responses without assuming the distributions for the errors. Empirical likelihood methods are proposed for constructing confidence intervals for the parameters of interest. Maximum empirical likelihood estimators, as well as their asymptotic properties, are obtained for both the cases that the threshold variable is known or not. A method to test the nonlinearity of the data is provided. As an illustration, we conduct a simulation study and empirical analysis of Pittsburgh crime data sets.

First-order random coefficients integer-valued threshold autoregressive processes

In this paper, we introduce a first-order random coefficient integer-valued threshold autoregressive process, which is based on binomial thinning. Basic probabilistic and statistical properties of this model are discussed. Conditional least squares and conditional maximum likelihood estimators are derived for both the cases that the threshold variable is known or not. The asymptotic properties of the estimators are established. Moreover, forecasting problem is addressed. Finally, some numerical results of the estimates and a real data example are presented.

An integer-valued threshold autoregressive process based on negative binomial thinning

In this paper, we introduce an integer-valued threshold autoregressive process, which is driven by independent negative-binomial distributed random variables and based on negative binomial thinning. Basic probabilistic and statistical properties of this model are discussed. Conditional least squares and conditional maximum likelihood estimators and corresponding iterative algorithms are investigated for both the cases that the threshold variable is known or not. Also, the asymptotic properties of the estimators are obtained. Finally, some numerical results of the estimates and a real data example are presented.

Univariate Conditional Distributions of an Open-Loop TAR Stochastic Process

<p>Clusters of large values are observed in sample paths of certain open-loop threshold autoregressive (TAR) stochastic processes. In order to characterize the stochastic mechanism that generates this empirical stylized fact, three types of marginal conditional distributions of the underlying stochastic process are analyzed in this paper. One allows us to find the conditional variance function that explains the aforementioned stylized fact. As a by-product, we are able to derive a sufficient condition to have asymptotic weak stationarity in an open-loop TAR stochastic process.</p>