mixest2 {dynmix}

Computes Mixture Estimation with Normal Regression Components.

Description

This function estimates recursively mixtures with normal regression components with a dynamic model of switching.

Usage

mixest2(y,x,mods=NULL,ftype=NULL,V=NULL,W=NULL,atype=NULL,Tvar=NULL)

Arguments

y

one column matrix of forecasted time-series, observations inserted rowwise

x

matrix of independent time-series (predictors), observations inserted rowwise

mods

see mixest1

ftype

optional, numeric indicating type of forecasting, 1 represents averaging forecasts from all components by the estimated weights (i.e., forecasting based on coefficients derived from the estimated mixture), 2 represents selecting the forecast given by the model with the highest weight, 3 represents selecting the forecast from the so-called median probability model (Barbieri and Berger, 2004), by default ftype=1 is taken

V

optional, numeric initial variance, the same for all components, by default V=1 is taken

W

optional, numeric initial value to be put in the diagonal matrix representing the covariance matrices for regression coefficients, the same for all components, by default W=1 is taken

atype

optional, numeric indicating approximation of pdfs, 0 represents quasi-Bayesian approach, 1 represents minimization of the Kerridge inaccuracy, by default atype=0 is taken

Tvar

optional, numeric indicating the number of first observations, when variance and covariance updating will not be performed, i.e., the initial values will be kept fixed, by default Tvar=30 is taken

Value

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

$y.hat

vector of predictions

$rvi

matrix of relative variable importances

$coef

matrix of regression coefficients corresponding to ftype method chosen

$weights

matrix of estimated weights of component models

$V

vector of updated variances from the selected models, consistent with ftype chosen

$R

matrix of updated diagonal of covariances corresponding to independent variables in regressions, consistent with ftype chosen

$components

matrix of mods

$parameters

character of parameters used in the model

Source

Nagy, I., Suzdaleva, E., Karny, M., Mlynarova, T., 2011, Bayesian estimation of dynamic finite mixtures. International Journal of Adaptive Control and Signal Processing 25, 765–787.

References

Barbieri, M. M., Berger, J. O., 2004, Optimal predictive model selection. The Annals of Statistics 32, 870–897.

Burnham, K. P., Anderson, D. R., 2002, Model Selection and Multimodel Inference, Springer.

Dedecius, K., 2010, Partial Forgetting in Bayesian Estimation, Czech Technical University in Prague.

Karny, M. (ed.), 2006, Optimized Bayesian Dynamic Advising, Springer.

Nagy, I., 2015, Mixture Models and Their Applications, Czech Technical University in Prague.

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

Quarteroni, A., Sacco, R., Saleri, F., 2007, Numerical Mathematics, Springer.

See Also

mixest1

Examples

data(oil)
m1 <- mixest2(y=oil[,1,drop=FALSE],x=oil[,-1,drop=FALSE],ftype=1,V=100,W=100)