create_stm: Create a Structural Topic Model

In NicolasJBM/lexR: Toolbox to manipulate and analyze texts

Description

Create a topic model from a document-to-term matrix.

Usage

create_stm(
  dtm,
  topic_nbr = 0,
  prevalence = NULL,
  content = NULL,
  init_type = "Spectral",
  seed = 1234,
  iterations = 50,
  tolerance = 0,
  keywords = NULL,
  baselab = "prob",
  frexweight = 0.6,
  min_gamma = 0.1,
  min_beta = 0.01,
  min_common = 1,
  verbose = TRUE
)

Arguments

dtm	Document-to-Term matrix.
topic_nbr	Numeric. Number of topics to model.
prevalence	Formula. See the stm function for detailed instructions.
content	Formula. See the stm function for detailed instructions.
init_type	Character string. See the stm function for detailed instructions.
seed	Numeric. Seed for initialization.
iterations	Integer. Maximum number of iteration.
tolerance	Numeric. See the stm function for detailed instructions.
keywords	Tibble. Two variables: one for the "document" and one of the related "keywords".
baselab	Character string. Whether labels should be based on "prob", "frex", "lift", or "score".
frexweight	Numeric. The weight on discriminance in term selection.
min_gamma	Numeric. Minimum strength of the relationship between document and topic.
min_beta	Numeric. Minimum strength of the relationship between term and topic.
min_common	Integer. Maximum number of topics allowed per document.
verbose	Logical. Should messages be printed.

Value

A list with the "topic_model", the "topic_term", the "topic_label", and the "topic_document".

Author(s)

Nicolas Mangin

See Also

stm

create_dtm

NicolasJBM/lexR documentation built on Feb. 4, 2021, 6:43 p.m.