cost_gaussian {bbemkr} | R Documentation |
Calculates the negative of log posterior, using the leave-one-out cross validated samples.
cost_gaussian(x, data_x, data_y, prior_p, prior_st)
x |
Log of square bandwidths |
data_x |
Regressors |
data_y |
Response variable |
prior_p |
A tuning parameter of the prior of error variance, following inverse gamma distribution |
prior_st |
Another tuning parameter of the prior of error variance, following inverse gamma distribution |
Bandwidth can be re-parameterized by a constant times optimal convergence rate, that is, h=c*n^{rate}. The prior of c^2 is
assumed to follow an inverse-gamma prior with hyperparameters prior_p = 2
and prior_st = 1
.
Value of the cost function
Han Lin Shang
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.
x = log(nrr(data_x, FALSE)^2) inicost = cost_gaussian(x, data_x = data_x, data_y = data_ynorm, prior_p = 2, prior_st = 1)