LAM-package.RdIncludes some procedures for latent variable modeling with a particular focus on multilevel data. The 'LAM' package contains mean and covariance structure modelling for multivariate normally distributed data (mlnormal(); Longford, 1987; <doi:10.1093/biomet/74.4.817>), a general Metropolis-Hastings algorithm (amh(); Roberts & Rosenthal, 2001, <doi:10.1214/ss/1015346320>) and penalized maximum likelihood estimation (pmle(); Cole, Chu & Greenland, 2014; <doi:10.1093/aje/kwt245>).
The LAM package contains the following main functions:
A general fitting method for mean and covariance structure for
multivariate normally distributed data is the mlnormal
function. Prior distributions or regularization methods (lasso penalties)
are also accommodated. Missing values on dependent variables can be
treated by applying the full information maximum likelihood method
implemented in this function.
A general (but experimental) Metropolis-Hastings sampler for Bayesian
analysis based on MCMC is implemented in the amh function.
Deterministic optimization of the posterior distribution (maximum
posterior estimation or penalized maximum likelihood estimation) can be
conduction with the pmle function which is based on
stats::optim.
Cole, S. R., Chu, H., & Greenland, S. (2013). Maximum likelihood, profile likelihood, and penalized likelihood: a primer. American Journal of Epidemiology, 179(2), 252-260. doi:10.1093/aje/kwt245
Longford, N. T. (1987). A fast scoring algorithm for maximum likelihood estimation in unbalanced mixed models with nested random effects. Biometrika, 74(4), 817-827. doi:10.1093/biomet/74.4.817
Roberts, G. O., & Rosenthal, J. S. (2001). Optimal scaling for various Metropolis-Hastings algorithms. Statistical Science, 16(4), 351-367. doi:10.1214/ss/1015346320
## > library(LAM)
## ## LAM 0.0-4 (2017-03-03 16:53:46)
##
## __ ______ __ __
## /\ \ /\ __ \ /\ "-./ \
## \ \ \____ \ \ __ \ \ \ \-./\ \
## \ \_____\ \ \_\ \_\ \ \_\ \ \_\
## \/_____/ \/_/\/_/ \/_/ \/_/
##