Skip to contents

Stability of topics across runs, by number of topics

Source: R/model_selection.R
plot_topic_stability.Rd

For each K (and corpus), the distribution over topics of their stability (topic_stability()): the mean cosine similarity between a topic of the reference run and its match in the other runs. A K whose topics mostly come back (similarity close to 1) gives results that do not depend on the luck of the initialisation (Greene, O'Callaghan and Cunningham 2014).

Usage

plot_topic_stability(
  x,
  measure = c("stability_cosine", "stability_jaccard"),
  base_size = 16,
  ...
)

Arguments

x

A topic_model_evaluation with stability results (metrics = "stability", n_runs >= 2).

measure

"stability_cosine" or "stability_jaccard".

base_size

Numeric scalar, base font size of the theme.

...

Not used: an argument renamed in the development version of mintR gives an error naming its replacement.

Value

A ggplot.

References

Greene, D., O'Callaghan, D. and Cunningham, P. (2014). How many topics? Stability analysis for topic models. Machine Learning and Knowledge Discovery in Databases (ECML PKDD 2014), 498-513.