A0767
Title: A comprehensive framework for investigating multiple latent class variables
Authors: Youngsun Kim - Korea University (Korea, South) [presenting]
Hwan Chung - Korea University (Korea, South)
Abstract: Latent class analysis (LCA) is a popular method for population segmentation, but it may not be sufficient for capturing complex population structures that require multiple latent class variables. Several approaches, such as Latent Transition Analysis (LTA), Latent Class Profile Analysis (LCPA), Joint Class Analysis (JLCA), and Joint Latent Class Profile Analysis (JLCPA), have been developed to explore the association among multiple latent class variables. A new framework is proposed, called the Structural Latent Class Model (SLCM), which integrates these existing LCA variants into a single framework by linking multiple latent class variables via transition matrices. A user-friendly R package for implementing the models has also been developed.
