LDA Variational EM Algorithm

This R package implements the variational expectation-maximization (VEM) algorithm for the latent Dirichlet allocation (LDA) model. This VEM algorithm is for the LDA full Bayesian model. Except for this additional feature, this implementation is quite similar to the original implementation of the LDA VEM algorithm by Dr. David Blei.

For package documentation run

help("ldavem")

in an R console. All major functions and datasets are documented and linked to the package index. Raw data files for each dataset are available in the data-raw folder. To load raw data see demo/load_raw_data.R .

To see all demo R scripts available in this package, run

demo(package="ldavem")

in an R console. Some scripts can be executed via running

demo(file-name, package="ldavem")

in an R console. The rest of them may require commandline arguments for execution. Please see the documentation provided in each script before execution.

Authors

Clint P. George (Please contact for questions and comments)

Dependencies

This package uses the following R packages, which are already included in this R package.

Rcpp
RcppArmadillo based on the Armadillo C++ package
lattice
Hyperparameter optimization is adapted from Blei (2004)'s implementation

Installation Guide

Download the package source from Git Download Link
Unzip the dowloaded file and rename the folder ldavem-master to ldavem
To install ldavem run R CMD INSTALL ldavem on the commandline
To uninstall ldavem run R CMD REMOVE ldavem on the commandline

References

Blei, D. M., Ng, A. Y. and Jordan, M. I. (2003). Latent Dirichlet allocation. Journal of Machine Learning Research 3 993-1022.
Blei, D. M. (2004). C implementation of variational EM for latent Dirichlet allocation (LDA)

Acknowledgements

Clint is supported by the NIH Grant #7 R21 GM101719-03 and the University of Florida Informatics Institute.

I would like to thank Dr. George Michaildis and Wei Xia for the valuable discussions and critics that helped the development.