Choosing the Number of Topics in LDA Models – A Monte Carlo Comparison of Selection Criteria

Victor Bystrov; Viktoriia Naboka-Krell; Anna Staszewska-Bystrova; Peter Winker

Selecting the number of topics in Latent Dirichlet Allocation (LDA) models is considered to be a difficult task, for which various approaches have been proposed. In this paper the performance of the recently developed singular Bayesian information criterion (sBIC) is evaluated and compared to the performance of alternative model selection criteria. The sBIC is a generalization of the standard BIC that can be applied to singular statistical models. The comparison is based on Monte Carlo simulations and carried out for several alternative settings, varying with respect to the number of topics, the number of documents and the size of documents in the corpora. Performance is measured using different criteria which take into account the correct number of topics, but also whether the relevant topics from the considered data generation processes (DGPs) are revealed. Practical recommendations for LDA model selection in applications are derived.

Choosing the Number of Topics in LDA Models – A Monte Carlo Comparison of Selection Criteria

Abstract