Thematic structure of AI-supported simulation-based learning: LDA topic modeling data and code

Data and code supporting an LDA topic modeling study of 363 Scopus-indexed publications (2010 - June 2026) on AI-supported simulation-based learning. The deposit contains the DOIs of the analysed corpus, the document-topic matrix (theta), topic-term distributions with topic labels and meta-theme groupings, the topic-number search scores, model diagnostics, figures, a pyLDAvis visualisation, and the full analysis pipeline in Python. Scopus titles and abstracts are licensed content and are not redistributed; DOIs are provided so that the corpus can be reconstructed by any Scopus subscriber. Model: Gensim LdaModel, K = 13, alpha = auto, eta = auto, 50 passes, 400 iterations, random_state = 42; coherence c_v = 0.2811. Author details are withheld while the associated manuscript is under anonymised peer review and will be added after acceptance. Files in code/ are MIT licensed; all other files are CC BY 4.0.