TY - JOUR
T1 - Latent class analysis of incomplete data via an entropy-based criterion
AU - Larose, Chantal
AU - Harel, Ofer
AU - Kordas, Katarzyna
AU - Dey, Dipak K.
N1 - Funding Information:
The research was partially supported by Award Number K01MH087219 from the National Institute of Mental Health , the University of Connecticut CLAS Graduate Fellowship , and the University of Connecticut Department of Statistics Elizabeth Macfarlane Fellowship . The authors are entirely responsible for the contents of this paper. The paper does not reflect the official views of the National Institute of Mental Health nor those of the National Institutes of Health.
Funding Information:
The computation was done partially on the Beowulf cluster of the University of Connecticut Department of Statistics. The cluster is partially financed by the NSF SCREMS (Scientific Computing Research Environments for the Mathematical Sciences) Grant Number 0723557 . All tables were formatted using the R package xtable (Dahl, 2014).
Funding Information:
The data collection for the application of our method was funded by a seed grant from the Children, Youth and Families Consortium at the Pennsylvania State University (Kordas, PI). Sincerest thanks to Dr. Elena Queirolo, Psychologist Graciela Ardoino, and Dr. Nelly Mañay for coordinating the study. In addition, many thanks to the Catholic University of Uruguay study research team for their data collection.
Publisher Copyright:
© 2016 Elsevier B.V.
PY - 2016/9/1
Y1 - 2016/9/1
N2 - Latent class analysis is used to group categorical data into classes via a probability model. Model selection criteria then judge how well the model fits the data. When addressing incomplete data, the current methodology restricts the imputation to a single, pre-specified number of classes. We seek to develop an entropy-based model selection criterion that does not restrict the imputation to one number of clusters. Simulations show the new criterion performing well against the current standards of AIC and BIC, while a family studies application demonstrates how the criterion provides more detailed and useful results than AIC and BIC.
AB - Latent class analysis is used to group categorical data into classes via a probability model. Model selection criteria then judge how well the model fits the data. When addressing incomplete data, the current methodology restricts the imputation to a single, pre-specified number of classes. We seek to develop an entropy-based model selection criterion that does not restrict the imputation to one number of clusters. Simulations show the new criterion performing well against the current standards of AIC and BIC, while a family studies application demonstrates how the criterion provides more detailed and useful results than AIC and BIC.
UR - http://www.scopus.com/inward/record.url?scp=84975086199&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84975086199&partnerID=8YFLogxK
U2 - 10.1016/j.stamet.2016.04.004
DO - 10.1016/j.stamet.2016.04.004
M3 - Article
AN - SCOPUS:84975086199
SN - 1572-3127
VL - 32
SP - 107
EP - 121
JO - Statistical Methodology
JF - Statistical Methodology
ER -