Using the Reversible Jump MCMC Procedure for Identifying and Estimating Univariate TAR Models

One way that has been used for identifying and estimating threshold autoregressive (TAR) models for nonlinear time series follows the Markov chain Monte Carlo (MCMC) approach via the Gibbs sampler. This route has major computational difficulties, specifically, in getting convergence to the parameter distributions. In this article, a new procedure for identifying a TAR model and for estimating its parameters is developed by following the reversible jump MCMC procedure. It is found that the proposed procedure conveys a Markov chain with convergence properties.     

A New Acceptance Sampling Policy Based on Number of Successive Conforming Items

In an acceptance-sampling plan, where items of an incoming batch of products are inspected one by one, if the number of conforming items between successive non conforming items falls below a lower control threshold, the batch is rejected. If it falls above an upper control threshold, the batch is accepted, and if it lies within the thresholds then the process of inspecting the items continues. The purpose of this article is to develop an optimization model to determine the optimum values of the thresholds such that constraints on the probability of Type I and Type II errors are satisfied. This article starts by developing a Markovian model to derive the expected total cost of the inspection problem containing the costs of acceptance, rejection, and inspection. Then, the optimum values of the thresholds are selected in order to minimize the expected cost. To demonstrate the application of the proposed methodology, perform sensitivity analysis, and compare the performance of the proposed procedure to the one of another method, a numerical example is given at the end and the results are reported.     

门限自回归下的单位根检验可靠性研究

在非对称的门限自回归模型下,由于传统单位根检验式的误设,会导致单位根检验势下降。本文通过一系列的Monte-Carlo模拟表明：非对称性对ADF和PP检验的检验势会产生较大影响,而对其他四种常用的单位根检验势产生的影响较小,也就是说,在非对称的门限自回归下,非对称性对退势单位根检验势产生的影响较小。模拟中也发现：NP单位根检验对TAR模型和持久性都具有稳健性。     

具有GARCH（1，1）-正态误差项的非对称单位根检验研究

摘　　要:理论研究表明许多经济变量呈现出非对称的门限自回归（TAR）或动态门限自回归（M-TAR）数据生成机制,因而非对称单位根检验就成为该领域的主要研究方向之一。本文对非对称单位根检验Enders-Granger方法在GARCH（1,1）-正态误差项下的检验水平与检验势作了系统的仿真研究。研究表明：GARCH（1,1）-正态误差项的TAR或M-TAR模型会对该方法的检验水平和检验势产生重要影响。     

A General Multivariate Threshold GARCH Model With Dynamic Conditional Correlations

We introduce a new multivariate GARCH model with multivariate thresholds in conditional correlations and develop a two-step estimation procedure that is feasible in large dimensional applications. Optimal threshold functions are estimated endogenously from the data and the model conditional covariance matrix is ensured to be positive definite. We study the empirical performance of our model in two applications using U.S. stock and bond market data. In both applications our model has, in terms of statistical and economic significance, higher forecasting power than several other multivariate GARCH models for conditional correlations.     

平稳阈值自回归下的伪回归研究

 研究表明相互独立的平稳阈值自回归(TAR)模型之间的回归存在伪回归,且伪回归的产生与样本容量和随机干扰项的分布无关。通过一系列的MC模拟,不仅证实了理论结论,而且模拟结果还表明当持久性相同时,两机制TAR回归模型比三机制TAR回归模型具有更大的拒绝率,原因在于两机制TAR下,OLS法估计得到的标准误具有更厚的左尾。此外在模拟中也发现当随机干扰项服从TAR模型时,Newey-West(1987)的一致异方差估计法是不适用的。     

Common threshold in quantile regressions with an application to pricing for reputation

The paper develops a systematic estimation and inference procedure for quantile regression models where there may exist a common threshold effect across different quantile indices. We first propose a sup-Wald test for the existence of a threshold effect, and then study the asymptotic properties of the estimators in a threshold quantile regression model under the shrinking threshold effect framework. We consider several tests for the presence of a common threshold value across different quantile indices and obtain their limiting distributions. We apply our methodology to study the pricing strategy for reputation through the use of a data set from Taobao.com. In our economic model, an online seller maximizes the sum of the profit from current sales and the possible future gain from a targeted higher reputation level. We show that the model can predict a jump in optimal pricing behavior, which is considered as “reputation effect” in this paper. The use of threshold quantile regression model allows us to identify and explore the reputation effect and its heterogeneity in data. We find both reputation effects and common thresholds for a range of quantile indices in seller’s pricing strategy in our application.     

Double threshold autoregressive conditionally heteroscedastic model building by genetic algorithms

Asymmetric behaviour in both mean and variance is often observed in real time series. The approach we adopt is based on double threshold autoregressive conditionally heteroscedastic (DTARCH) model with normal innovations. This model allows threshold nonlinearity in mean and volatility to be modelled as a result of the impact of lagged changes in assets and squared shocks, respectively. A methodology for building DTARCH models is proposed based on genetic algorithms (GAs). The most important structural parameters, that is regimes and thresholds, are searched for by GAs, while the remaining structural parameters, that is the delay parameters and models orders, vary in some pre-specified intervals and are determined using exhaustive search and an Asymptotic Information Criterion (AIC) like criterion. For each structural parameters trial set, a DTARCH model is fitted that maximizes the (penalized) likelihood (AIC criterion). For this purpose the iteratively weighted least squares algorithm is used. Then the best model according to the AIC criterion is chosen. Extension to the double threshold generalized ARCH (DTGARCH) model is also considered. The proposed methodology is checked using both simulated and market index data. Our findings show that our GAs-based procedure yields results that comparable to that reported in the literature and concerned with real time series. As far as artificial time series are considered, the proposed procedure seems to be able to fit the data quite well. In particular, a comparison is performed between the present procedure and the method proposed by Tsay [Tsay, R.S., 1989, Testing and modeling threshold autoregressive processes. Journal of the American Statistical Association, Theory and Methods, 84, 231–240.] for estimating the delay parameter. The former almost always yields better results than the latter. However, adopting Tsay's procedure as a preliminary stage for finding the appropriate delay parameter may save computational time specially if the delay parameter may vary in a large interval.     

Nonlinearity testing and modeling for threshold moving average models

In this paper, we suggest a simple test and an easily applicable modeling procedure for threshold moving average (TMA) models. Firstly, based on the fitted residuals by maximum likelihood estimate (MLE) for MA models, we construct a simple statistic, which is obtained by linear arrange regression and follows F-distribution approximately, to test for threshold nonlinearity and specify the threshold variables. And then, we use some scatterplots to identify the number and locations of the potential thresholds. Finally, with the statistic and Akaike information criterion, we propose the procedure to build TMA models. Both the power of test statistic and the convenience of modeling procedure can work very well demonstrated by simulation experiments and the application to a real example.     

Do Flow Rates Respond Asymmetrically to Water Level? Evidence from the Edwards Aquifer

This research examines the time series relationship between the Comal Springs flow rate and the water level in the Edwards Aquifer (Well J-17). The empirical methodology utilizes threshold autoregression (TAR) and momentum-TAR models that allow for asymmetry in responses and adjustments to a disequilibrium in the long-run cointegrating relationship. Based on the results, an asymmetric error-correction model (AECM) is proposed to characterize the short-run and long-run dynamic relationship between spring flow and water level. The results have implications for the management of water resources, water demand, and ecosystems.     

Optimal investment-consumption-insurance strategy in a continuous-time self-exciting threshold model

In this paper, we consider an optimal investment-consumption-insurance purchase problem for a wage earner. We assume that the price of the risky asset is governed by a continuous-time, finite state self-exciting threshold model. In this model, the state space of the price of the risky asset is partitioned by a set of thresholds and the parameters depend on the region which the current value of the price falls in. The wage earner’s objective is to find the optimal investment-consumption-insurance strategy that maximizes the expected discounted utilities. The optimal strategy for power utility function is derived by the martingale approach and the dynamic programming approach. Numerical examples are also provided to illustrate the effect of the thresholds.     

Second-order refined peaks-over-threshold modelling for heavy-tailed distributions

Modelling excesses over a high threshold using the Pareto or generalized Pareto distribution (PD/GPD) is the most popular approach in extreme value statistics. This method typically requires high thresholds in order for the (G)PD to fit well and in such a case applies only to a small upper fraction of the data. The extension of the (G)PD proposed in this paper is able to describe the excess distribution for lower thresholds in case of heavy-tailed distributions. This yields a statistical model that can be fitted to a larger portion of the data. Moreover, estimates of tail parameters display stability for a larger range of thresholds. Our findings are supported by asymptotic results, simulations and a case study.     

A boosting inspired personalized threshold method for sepsis screening

Sepsis is one of the biggest risks to patient safety, with a natural mortality rate between 25% and 50%. It is difficult to diagnose, and no validated standard for diagnosis currently exists. A commonly used scoring criteria is the quick sequential organ failure assessment (qSOFA). It demonstrates very low specificity in ICU populations, however. We develop a method to personalize thresholds in qSOFA that incorporates easily to measure patient baseline characteristics. We compare the personalized threshold method to qSOFA, five previously published methods that obtain an optimal constant threshold for a single biomarker, and to the machine learning algorithms based on logistic regression and AdaBoosting using patient data in the MIMIC-III database. The personalized threshold method achieves higher accuracy than qSOFA and the five published methods and has comparable performance to machine learning methods. Personalized thresholds, however, are much easier to adopt in real-life monitoring than machine learning methods as they are computed once for a patient and used in the same way as qSOFA, whereas the machine learning methods are hard to implement and interpret.     

A new estimation method for continuous threshold expectile model

The continuous threshold expectile regression model could capture the effect of a covariate on the response variable with two different straight lines, while intersecting an unknown threshold needed be estimated. This article proposes a new estimation method via a linearization technique to estimate the regression coefficients and the threshold simultaneously. Statistical inferences of the proposed estimators are easily derived from the existing theory. Moreover, the estimation procedure is readily implemented by the current software. Simulation studies and an application on GDP per capita and quality of electricity supply data illustrate the proposed method.     

Uneven two-sided power distributions with applications in econometric models

In this paper we propose and analyze a bounded density function with a jump discontinuity at a threshold. Its properties are presented and a maximum likelihood estimation (MLE) procedure for the threshold location and jump size is developed. The distribution seems be appropriate in the context of financial engineering, production analysis, standard auction models and the equilibrium job search problem. An example of the MLE procedure is given utilizing an i.i.d. sample of standardized log differences of bi-monthly US Certificate Deposit interest rates for the period from 1966-2002. The corresponding time series was constructed using an Auto-Regressive Conditional Heteroscedastic (ARCH) model.     

Threshold selection for regional peaks-over-threshold data

A hurdle in the peaks-over-threshold approach for analyzing extreme values is the selection of the threshold. A method is developed to reduce this obstacle in the presence of multiple, similar data samples. This is for instance the case in many environmental applications. The idea is to combine threshold selection methods into a regional method. Regionalized versions of the threshold stability and the mean excess plot are presented as graphical tools for threshold selection. Moreover, quantitative approaches based on the bootstrap distribution of the spatially averaged Kolmogorov–Smirnov and Anderson–Darling test statistics are introduced. It is demonstrated that the proposed regional method leads to an increased sensitivity for too low thresholds, compared to methods that do not take into account the regional information. The approach can be used for a wide range of univariate threshold selection methods. We test the methods using simulated data and present an application to rainfall data from the Dutch water board Vallei en Veluwe.     

Optimal Threshold from ROC and CAP Curves

Receiver Operating Characteristic (ROC) and Cumulative Accuracy Profile (CAP) curves are used to assess the discriminatory power of different credit-rating approaches. The thresholds of optimal classification accuracy on an ROC curve and of maximal profit on a CAP curve can be found by using iso-performance tangent lines, which are based on the standard notion of accuracy. In this article, we propose another accuracy measure called the true rate. Using this rate, one can obtain alternative optimal thresholds on both ROC and CAP curves. For most real populations of borrowers, the number of the defaults is much less than that of the non defaults, and in such cases using the true rate may be more efficient than using the accuracy rate in terms of cost functions. Moreover, it is shown that both alternative optimal thresholds by using the true rate are the identical, and this single threshold coincides with the score corresponding to Kolmogorov–Smirnov statistic used to test the homogeneous distribution functions of the defaults and non defaults, whereas the optimal threshold by using the accuracy does not the same as the score corresponding to Kolmogorov–Smirnov statistic. These facts are explored with some simulation and illustrative examples.     

非对称单位根检验和非参数方差比

基于三种退势方法较详细研究了方差比检验在非对称单位根检验中的适用性,并通过MC模拟揭示了其检验势性质。结果表明：在不含趋势项的TAR下,两机制TAR数据落在第一机制的比率是影响方差比检验势的重要因素,且比率越高检验势也越高;三机制TAR中落在中间机制的数据比率会影响检验势,随着比率增加检验势呈下降趋势,但程度不大。在含趋势的TAR下,由于趋势项在数据生成过程中具有支配作用,各种检验势会随着趋势设定的不同而不同。数据在不同机制之间的转换概率越高,则ROLS和RDM退势较OLS退势具有明显优势。     

Estimation of extreme survival probabilities with cox model

We propose an extension of the regular Cox's proportional hazards model which allows the estimation of the probabilities of rare events. It is known that when the data are heavily censored, the estimation of the tail of the survival distribution is not reliable. To improve the estimate of the baseline survival function in the range of the largest observed data and to extend it outside, we adjust the tail of the baseline distribution beyond some threshold by an extreme value model under appropriate assumptions. The survival distributions conditioned to the covariates are easily computed from the baseline. A procedure allowing an automatic choice of the threshold and an aggregated estimate of the survival probabilities are also proposed. The performance is studied by simulations and an application on two data sets is given.     

Testing a linear ARMA model against threshold-ARMA models: A Bayesian approach

We introduce a Bayesian approach to test linear autoregressive moving-average (ARMA) models against threshold autoregressive moving-average (TARMA) models. First, the marginal posterior densities of all parameters, including the threshold and delay, of a TARMA model are obtained by using Gibbs sampler with Metropolis–Hastings algorithm. Second, reversible-jump Markov chain Monte Carlo (RJMCMC) method is adopted to calculate the posterior probabilities for ARMA and TARMA models: Posterior evidence in favor of TARMA models indicates threshold nonlinearity. Finally, based on RJMCMC scheme and Akaike information criterion (AIC) or Bayesian information criterion (BIC), the procedure for modeling TARMA models is exploited. Simulation experiments and a real data example show that our method works well for distinguishing an ARMA from a TARMA model and for building TARMA models.