Level: Advanced
【Presenter】
National University Corporation Hokkaido University Hokkaido University Hospital
Medical and Health Science Research and Development Organization
Data Science Center Director Yoichi Ito
[Presentation summary]
Latent variable model analysis is an analysis method that defines some property behind the object of analysis as a latent variable that cannot be directly observed, and assumes that the latent variable can be observed indirectly. Latent variable models are classified into four types depending on the combination of observed variables and latent variable measures. When the observed variable is a continuous scale and the latent variable is also a continuous scale, factor analysis is used. When the observed variable is a continuous scale and the latent variable is a discrete scale, it is a finite mixture model. When the observed variable is a discrete scale, it is a finite mixture model. When the latent variable is on a continuous scale, it is an item response theory model, and when the observed variable is on a discrete scale and the latent variable is also on a discrete scale, it is called latent class analysis. Of these four analysis methods, latent class analysis is not implemented in SAS and can only be performed in JMP Pro. In this presentation, we will explain the relationship between these four analysis methods and introduce latent variable model analysis using JMP.
[Presenter profile]
From May 2001, he became an assistant professor of biostatistics at the Graduate School of Medicine, the University of Tokyo. In 2008, he became a specially appointed lecturer at the Hokkaido University Graduate School of Medicine. In 2011, he became an associate professor. In 2018, he joined the Institute of Statistical Mathematics, Research Organization of Information and Systems. Became director of the Medical and Health Data Science Research Center. In 2020, he became Professor and Head of the Department of Biostatistics, Clinical Research and Development Center, Hokkaido University Hospital, and assumed his current position in 2021 with the establishment of the Medical and Health Science Research and Development Organization.
He serves as a statistical consultant to researchers conducting clinical research and as the person in charge of statistical analysis for investigator-initiated clinical trials. Additionally, as an expert member of the Pharmaceuticals and Medical Devices Agency (PMDA), he is involved in approximately 10 new drug reviews a year.
*We are currently checking whether the distributed materials can be made public.
