gibbs_admkr_erro: Estimating bandwidth of the kernel-form error density

In bbemkr: Bayesian bandwidth estimation for multivariate kernel regression with Gaussian error

Description

Implements the random-walk Metropolis algorithm to estimate the bandwidth of the kernel-form error density

Usage

gibbs_admkr_erro(xh, inicost, k, errorsizp, errorprob, data_x, data_y)  

Arguments

xh	Log of square bandwidth in the kernel-form error density
inicost	Cost value
k	Iteration number
errorsizp	Step size of random-walk Metropolis algorithm
errorprob	Optimal convergence rate for drawing single or multiple parameters
data_x	Regressors
data_y	Response variable

Details

1) The log bandwidths of the regressors are initialized using the normal reference rule of Silverman (1986).

2) Conditioning on the variance parameter of the error density, we implement random-walk Metropolis algorithm to update the bandwidths, in order to achieve the minimum cost value.

3) The bandwidth of the kernel-form error density can be directly sampled.

4) Iterate steps 2) and 3) until the cost value is minimized.

5) Check the convergence of the parameters by examining the simulation inefficient factor (sif) value. The smaller the sif value is, the better convergence of the parameters is.

Value

x	Estimated bandwidth of the kernel-form error density
cost	Cost value, that is negative of log posterior
accept_erro	Accept or reject. accept_erro = 1 indicates acceptance, while accept_erro = 0 indicates rejection.
errorsizp	Step size of the random-walk Metropolis algorithm

Author(s)

Han Lin Shang

References

X. Zhang and R. D. Brooks and M. L. King (2009) A Bayesian approach to bandwidth selection for multivariate kernel regression with an application to state-price density estimation, Journal of Econometrics, 153, 21-32.

B. W. Silverman (1986) Density Estimation for Statistics and Data Analysis. Chapman and Hall, New York.

See Also

mcmcrecord_admkr, logdensity_admkr, loglikelihood_admkr, logpriors_admkr, gibbs_admkr_nw

bbemkr documentation built on May 1, 2019, 10:53 p.m.