.png){kind=icon}
.png){kind=icon}
Return
Simultaneous Variable Selection and Estimation of Mixed Panel Count Data With Informative Observation Processes
.png){kind=icon}
PRE
AI
DOI：10.1002/sta4.70099.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Mixed panel count data occur in many studies, and many researchers have analysed them. However, there seems to be no established approach for their analysis with the focus on variable selection when involving an informative observation process. It is widely acknowledged that methods assuming noninformative observation times can yield biased results when this assumption is violated. To fill the gap, a joint modelling strategy is introduced. Specifically, we develop a penalized likelihood estimation method and an efficient EM algorithm for simultaneous variable selection and estimation. The oracle property of the suggested method is demonstrated, and a simulation study is conducted, revealing its effective performance in practice. In addition, an illustration using the Health and Retirement Study data is provided.
Keywords:
EM algorithm
latent variable
minimum information criterion
panel count data
variable selection
Journal
S
IF:
0.8
Papers:
58
Citations:
655
Organization
View more organizations