posterior samples of model parameters.
a matrix or data.frame of size m \times p (m: sample size, p: dimension of the parameters).
par
MCLL parameter estimates on the original scale.
H.prior
Hessian matrix of the prior evaluated at par.
alp
a real value between 0 and 1. α takes a value between 0 and 1, which is the nearest neighbor bandwidth
with the kth smallest distance d where k = lfloor n α
floor and d(x, x_{i}) = | x - x_{i} |
with the sample size n.
method
an optimization method to be used to find the coefficients of the polynomial approximation
to the log-posterior at the MCLL estimates par.
Options from optim are Nelder-Mead, BFGS, CG, L-BFGS-B, and SANN.
lower, upper
bounds on the variables for the L-BFGS-B method in optim.
control
a list of control parameters. See control options for optim.
Details
Standard error estimation in the Monte Carlo local likelihood method.
For details, see Section 3 in Jeon et al. (2012).
The posterior samples and paramter values should be on the real line (e.g., variance parameters should be in the log-scale).
Value
mcll_se returns a vector containing standard error estimates for the MCLL parameter estimates par.
Author(s)
Minjeong Jeon <jeon.117@osu.edu>
References
Jeon, M., Kaufman, C., and Rabe-Hesketh, S. (2014). Monte Carlo local likelihood
for approximate MLE of complex models. Under revision.