tvp.reg: Computes Time-Varying Parameters Regression.

In dynmix: Estimation of Dynamic Finite Mixtures

Computes Time-Varying Parameters Regression.

Description

This function estimates Time-Varying Parameters regression.

Usage

tvp.reg(y,x,lambda=NULL,kappa=NULL,V=NULL,W=NULL)

Arguments

y	one column matrix of forecasted time-series, observations inserted rowwise
x	matrix of independent time-series (predictors), observations inserted rowwise
lambda	optional, see mixest1
kappa	optional, see mixest1
V	optional, numeric initial variance, by default V=1 is taken
W	optional, numeric initial value to be put on diagonal of covariance matrix, by default W=1 is taken

Details

If lambda is specified, then the method described by Raftery et al. (2010) is used, with possible extentsion to the one described by Koop and Korobilis (2012). Otherwise, the Kalman filter described as by Nagy and Suzdaleva (2013) is used.

Value

object of class tvpreg, i.e., list of

$y.hat	vector of predictions
$coef	matrix of regression coefficients
$R	matrix of diagonals of covariances corresponding to independent variables in regressions
$V	vector of outcome variances

References

Koop, G., Korobilis, D., 2012, Forecasting inflation using Dynamic Model Averaging. International Economic Review 53, 867–886.

Nagy, I., Suzdaleva, E., 2017, Algorithms and Programs of Dynamic Mixture Estimation, Springer.

Raftery, A. E., Karny, M., Ettler, P., 2010, Online prediction under model uncertainty via Dynamic Model Averaging: Application to a cold rolling mill. Technometrics 52, 52–66.

Examples

data(oil)
t1 <- tvp.reg(y=oil[,1,drop=FALSE],x=oil[,-1,drop=FALSE],lambda=0.99,V=100,W=100)
t2 <- tvp.reg(y=oil[,1,drop=FALSE],x=oil[,-1,drop=FALSE],V=100,W=100)

dynmix documentation built on July 9, 2023, 7:22 p.m.