Searches for the optimal number of topics (K) using stm::searchK. Produces diagnostic plots to help select the best K value.
Usage
find_optimal_k(
dfm_object,
topic_range,
max.em.its = 75,
emtol = 1e-04,
cores = 1,
categorical_var = NULL,
continuous_var = NULL,
height = 600,
width = 800,
verbose = TRUE,
...
)Arguments
- dfm_object
A quanteda dfm object to be used for topic modeling.
- topic_range
A vector of K values to test (e.g., 2:10).
- max.em.its
Maximum number of EM iterations (default: 75).
- emtol
Convergence tolerance for EM algorithm (default: 1e-04). Higher values (e.g., 1e-03) speed up fitting but may reduce precision.
- cores
Number of CPU cores to use for parallel processing (default: 1). Set to higher values for faster searchK on multi-core systems.
- categorical_var
Optional categorical variable(s) for prevalence.
- continuous_var
Optional continuous variable(s) for prevalence.
- height
Plot height in pixels (default: 600).
- width
Plot width in pixels (default: 800).
- verbose
Logical indicating whether to print progress (default: TRUE).
- ...
Additional arguments passed to stm::searchK.
See also
plot_quality_metrics() to visualize topic-count diagnostics; stm::stm() to fit the chosen model; fit_embedding_model() for an embedding-based alternative to STM