Nonlinear Time Series Models with Regime Switching
Additive nonlinear autoregressive model
Forecasting accuracy measures.
addRegime test
Long-term mean of an AR(p) process
Bivariate time series plots
Trivariate time series plots
Interactive trivariate time series plots
Available models
Test of unit root against SETAR alternative
Characteristic roots of the AR coefficients
Extract cointegration parameters A, B and PI
computeGradient
US unemployment series used in Caner and Hansen (2001)
delta test of linearity
delta test of conditional independence
Forecast Error Variance Decomposition
fitted method for objects of class nlVar, i.e. VAR and VECM models.
Extract threshold(s) coefficient
Generalized Impulse response Function (GIRF)
Impulse response function
isLinear
Test of unit root against SETAR alternative with
Selection of the lag with Information criterion.
Linear AutoRegressive models
Multivariate linear models: VAR and VECM
Locally linear model
Extract Log-Likelihood
Logistic Smooth Transition AutoRegressive model
Specification of the threshold search
Mean Absolute Percent Error
Mean Square Error
NLAR methods
Non-linear time series model, base class definition
NLAR common structure
Neural Network nonlinear autoregressive model
oneStep
Plotting methods for SETAR and LSTAR subclasses
Plot the Error Correct Term (ECT) response
Predict method for objects of class ‘nlar’ .
Predict method for objects of class ‘VAR’ , ‘VECM’ or ‘TVAR’
Rolling forecasts
Selection of the cointegrating rank with Information criterion.
Test of the cointegrating rank
Objects exported from other packages
Extract a variable showing the regime
Resampling schemes
Residual variance
Automatic selection of model hyper-parameters
Automatic selection of SETAR hyper-parameters
Self Threshold Autoregressive model
Simulation and bootstrap of Threshold Autoregressive model (SETAR)
Test of linearity against threshold (SETAR)
sigmoid functions
STAR model
Latex representation of fitted setar models
Getting started with the tsDyn package
Test of linearity
Multivariate Threshold Vector Autoregressive model
Simulation of a multivariate Threshold Autoregressive model (TVAR)
Test of linear cointegration vs threshold cointegration
Threshold Vector Error Correction model (VECM)
No cointegration vs threshold cointegration test
Simulation and bootstrap a VECM or bivariate TVECM
Simulate or bootstrap a VAR model
VAR representation
Estimation of Vector error correction model (VECM)
Virtual VECM model
Implements nonlinear autoregressive (AR) time series models. For univariate series, a non-parametric approach is available through additive nonlinear AR. Parametric modeling and testing for regime switching dynamics is available when the transition is either direct (TAR: threshold AR) or smooth (STAR: smooth transition AR, LSTAR). For multivariate series, one can estimate a range of TVAR or threshold cointegration TVECM models with two or three regimes. Tests can be conducted for TVAR as well as for TVECM (Hansen and Seo 2002 and Seo 2006).
Useful links