Predictive Likelihood calculation using Importance Sampling and mixture of Student-t densities as candidate

Share:

Description

Calculation of predictive likelihoods using Importance Sampling, given subsample and full data sample and Mixture of Student-t candidate density. Predictive likelihood is calculated using the marginal likelihood from full sample and subsample. See MargLik.

Usage

1
PredLik(N=1e4,mit.fs,mit.ss,KERNEL,data.fs,data.ss,...)

Arguments

N

integer > 100 number of draws for Importance Sampling

mit.fs

Mixture of Student-t density for the full sample, list describing the mixture of Student-t. See isMit. The mixture density can be obtained from MitISEM or SeqMitISEM

mit.ss

Mixture of Student-t density for subsample. Must be defined as mit.fs.

KERNEL

Function/posterior to be approximated. data and log arguments must exist. The function must return log-density if log=TRUE. All data should be loaded in argument data

data.fs

Full data, vector (length T1) or matrix (size T1xm) with data values, T1 observations and m data series.

data.ss

Sample of data, vector (length T2) or matrix (size T2xm) with data values, T2 observations and m data series. T2 <T1.

...

other arguments to be passed to KERNEL

Details

Argument KERNEL

Value

list containing:

PL

Predictive likelihood x 10^{scale}

scale

integer > 0 providing the scaling for predictive likelihood. (scaling may be necessary for numerical accuracy)

References

Eklund, J. and Karlsson, S. (2007). Forecast combination and model averaging using predictive measures. Econometric Reviews, 26, 329-363.

Min, C. and Zellner, A. (1993). Bayesian and non-Bayesian methods for combining models and forecasts with applications to forecasting international growth rates. Journal of Econometrics, 56, 89-118.

See Also

isMit,MargLik,MitISEM,SeqMitISEM