Introduction to Latent Class Analysis with an Application to Vitiligo
Latent Class Analysis (LCA) is a type of finite (discrete) mixture model that is used to discover the distribution of unobserved, underlying sub-classes within a data set. This differs from Latent Profile Analysis (LPA) in that LCA expects a binary outcome, whereas LPA expects a continuous result. This analysis is performed in R using the poLCA package.
This project analyses and confirms the findings of a previous study (Ezzedine et. al; please see document and/or presentation for citation) on the skin disease vitiligo with a new and expanded data set. With relatively few variables compared to the Ezzidine study, we are still able to confirm the probable existence of two sub-classes of the disease, coined "pre-pubertal" and "post-pubertal" vitiligo.
Wyatt "Miller" Miller
Data ScienceMillerUniversity of Colorado DenverData Science, Case Study, Statistics, Project Management, Mathematics, R, poLCA, LCA, Coding, Proofcraft, All ProjectsComment