Latent class modeling: basics, extensions, and recent advances
- Name of instructor:
Prof. Daniel Oberski, Utrecht University, The Netherlands. - Short description:
Latent class (finite mixture) modeling (LCM) is used in a wide variety of disciplines, including computer science, biology, medicine, genetics, chemistry, sociology, psychology, marketing, and economics. LCM describes multivariate data as arising from unobserved (“latent”, “hidden”) categorical (discrete) variables, and allows the practictioner to recover descriptions of these hidden variables, in fer relationships among them, and interpret their meaning in terms of the observed data. - Introductory background:
Vermunt, J.K. (2010). Latent class models. In: P. Peterson, E. Baker, B. McGaw, (eds.),
International Encyclopedia of Education, Volume 7, 238-244. Oxford: Elsevier.
URL: http://members.home.nl/jeroenvermunt/iee2008.pdf - Tentative schedule
1. Basic principles of LCM: latent structure model, identification, estimation
, interpretation [45 min]
2. Evaluating model fit of LCM’s: selecting the number of classes, local fit,
sensitivity to assumptions [45 min]
3. Break [15 min]
4. Special models: local dependence models, multiple latent variables, Hidden Markov Models and extensions, multilevel latent class models [45 min].