MAP estimation of topic models in R. It implements estimation via posterior maximization for latent topic models in text analysis, as described in "On Estimation and Selection for Topic Models". These routines were previously part of textir, but have been spun off since version 2.0 of that package. This maptpx package is no longer actively maintained; instead, we are focusing on developing faster distributed factor models based on distributed multinomial regression, as implemented in the distrom package.
If you want to take advantage of openmp parallelization, uncomment the relevant flags in src/MAKEVARS before compiling.