computes stm model diagnostics

get_diag() is a helper function to compute average and median semanticCoherence and exclusivity for a number of stm models. The function does not work for models with content covariates.

get_diag(models, outobj)

Arguments

models	A list of stm models.
outobj	The out object containing documents for all stm models.

Value

Returns model diagnostics in a data frame.

Examples

library(stm)
#> stm v1.3.6 successfully loaded. See ?stm for help. 
#>  Papers, resources, and other materials at structuraltopicmodel.com
library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union
library(ggplot2)
library(quanteda)
#> Package version: 2.1.2
#> Parallel computing: 2 of 8 threads used.
#> See https://quanteda.io for tutorials and examples.
#> 
#> Attaching package: 'quanteda'
#> The following object is masked from 'package:utils':
#> 
#>     View

# prepare data
data <- corpus(gadarian, text_field = 'open.ended.response')
docvars(data)$text <- as.character(data)
data <- dfm(data, stem = TRUE, remove = stopwords('english'),
            remove_punct = TRUE)
out <- convert(data, to = 'stm')

# fit models
gadarian_3 <- stm(documents = out$documents,
                  vocab = out$vocab,
                  data = out$meta,
                  prevalence = ~ treatment + s(pid_rep),
                  K = 3,
                  max.em.its = 1, # reduce computation time for example
                  verbose = FALSE)

gadarian_5 <- stm(documents = out$documents,
                  vocab = out$vocab,
                  data = out$meta,
                  prevalence = ~ treatment + s(pid_rep),
                  K = 5,
                  max.em.its = 1, # reduce computation time for example
                  verbose = FALSE)

# get diagnostics
diag <- get_diag(models = list(
                 model_3 = gadarian_3,
                 model_5 = gadarian_5),
                 outobj = out)
# \dontrun{
# plot diagnostics
diag %>%
ggplot(aes(x = coherence, y = exclusivity, color = statistic))  +
  geom_text(aes(label = name), nudge_x = 5) + geom_point() +
  labs(x = 'Semantic Coherence', y = 'Exclusivity') + theme_light()
# }