Estimation of the Structural Topic Model R package stm version 1.3.8 CRAN Task View: NaturalLanguageProcessing (https://CRAN.R-project.org/view=NaturalLanguageProcessing) The Structural Topic Model (STM) allows researchers to estimate topic models with document-level covariates. The package also includes tools for model selection, visualization, and estimation of topic-covariate regressions. Methods developed in Roberts et. al. (2014) <doi:10.1111/ajps.12103> and Roberts et. al. (2016) <doi:10.1080/01621459.2016.1141684>. Vignette is Roberts et. al. (2019) <doi:10.18637/jss.v091.i02>. Software Depends: R (>= 3.5.0), methods Imports: Rcpp (>= 0.11.3), data.table, glmnet, grDevices, graphics, lda, Matrix, matrixStats, parallel, quadprog, quanteda, slam, splines, stats, stringr, utils LinkingTo: Rcpp, RcppArmadillo Suggests: clue, geometry, huge, hunspell, igraph, LDAvis, KernSmooth, NLP, rsvd, Rtsne, SnowballC, spelling, testthat, tm (>= 0.6), wordcloud Brandon Stewart <bms4@princeton.edu> Comprehensive R Archive Network (CRAN) Margaret Roberts [aut], Brandon Stewart [aut, cre], Dustin Tingley [aut], Kenneth Benoit [ctb] MIT + file LICENSE (https://CRAN.R-project.org/package=stm/LICENSE) 2025-09-03 application/tgz https://CRAN.R-project.org/package=stm doi:10.32614/CRAN.package.stm en-US