sageLabels:
Displays verbose labels that describe topics and topic-covariate groups in depth.
Description
For each topic or, when there is a covariate at the bottom of the model, for each topic-covariate
group, sageLabels provides a list of the highest marginal probability words, the highest marginal
FREX words, the highest marginal lift words, and the highest marginal
score words, where marginal means it is summing over all potential covariates. It also
provides each topic's Kappa (words associated with each topic) and
baselined Kappa (baseline word distribution).
Usage
sageLabels(model, n=7)
Arguments
model
A fitted STM model object.
n
The number of words to print per topic/topic-covariate set. Default is 7.
Value
marginal
A list of matrices, containing the high-probability labels, FREX labels, lift labels,
and high scoring words.
K
The number of topics in the STM.
covnames
Names of the covariate values used in the STM.
kappa
Words associated with topics, covariates, and
topic/covariate interactions.
kappa.m
Baseline word distribution.
n
The n parameter passed by the user to this function; number of words per topic
or topic-covariate pair (when covariates are used on the bottom of the model)
cov.betas
Covariate-specific beta matrices, listing for each covariate a matrix of highest-probability,
FREX, lift, and high scoring words. Note that the actual vocabulary has been substituted
for word indices.
Details
This can be used as an more detailed alternative to labelTopics.