Skew-t linear mixed-effects models for longitudinal data
Author
Christian Eduardo Galarza Morales
Co-author
-
Victor Hugo Lachos Davila
-
Larissa Avila Matos
-
Katherine Loor
-
-
-
Conference
65th ISI World Statistics Congress 2025
Format: IPS Abstract - WSC 2025
Keywords: censored and skewed data, heavy-tailed, linear-mixed-model, robust
Tuesday 7 October 2 p.m. - 3:40 p.m. (Europe/Amsterdam)
Abstract
We propose a linear mixed-effects model for censored data, incorporating skewed and heavy-tailed Student-t random effects and symmetric heavy-tailed errors. Parameter estimation is conducted via the ECME algorithm, with standard errors approximated using the empirical information matrix. Simulation studies confirm the model's robustness and its performance is illustrated with a real dataset, showing improved fit over traditional methods.