Online Program Home
My Program

Abstract Details

Activity Number: 434 - SPEED: Classification and Data Science
Type: Contributed
Date/Time: Tuesday, July 31, 2018 : 2:00 PM to 2:45 PM
Sponsor: Section on Statistical Learning and Data Science
Abstract #332967
Title: Assessing Divide-and-Conquer Latent Class Analysis
Author(s): Qiao Ma* and Meimeizi Zhu and Edward Mulrow
Companies: NORC at the University of Chicago and NORC at the University of Chicago and NORC at the University of Chicago
Keywords: Latent class analysis; Big data; Multivariate Categorical data; Statistical methods; Structural equation modeling

Latent class analysis (LCA) is generally used to construct unobserved classes from observed indicator variables. Fitting an LCA to a large dataset is challenging. Abarda et al. (2017) proposed a Divide-and-Conquer approach for LCA when dealing with big data. Divide-and-Conquer partitions a data set into multiple subsets with equal size, and fits a model to each subset. We assess LCA results in such a setting by fitting LCA models with and without the Divide-and-Conquer approach, and measure the relationships between the Bayesian Information Criteria (BIC) for the whole dataset and its subsets.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2018 program