lda: LDA Model Fitting

In wrathematics/yalda: Yet Another LDA package

Description

This routine uses the plda application from Google to fit a Latent Dirichlet Allocation model.

Currently IO is required "both ways".

Usage

  lda(infile, outfile, compute.loglik=TRUE, num.topics=10, alpha=50/num.topics, 
      beta=0.01, niter=15, burnin=10, verbose=FALSE)

Arguments

infile	A string pointing to the input dataset.
outfile	A string pointing to the output data file.
compute.loglik	logical; controls whether the model log-likelihood should be returned.
num.topics	integer; determines the number of topics in the topic model.
alpha	numeric;
beta	numeric;
niter	integer; the total number of iterations to perform in fitting the model.
burnin	The number of burnin iterations (for the Gibbs sampler). Must be no greater than niter

.

verbose

logical; controls if "Iteration <iter>" should be printed at each iteration.

wrathematics/yalda documentation built on May 4, 2019, 10:54 a.m.