First order methods for non-Gaussian nonlinear mixed-effects models in Pumas
When: Tuesday March 23 2021, 9:30 AM - 10:30 AM Eastern Time (US)
Traditionally, the first-order methods for mixed-effects models have been associated with the Gaussian error models. However, when viewed as a Laplace approximation, the first-order methods apply beyond the Gaussian error model. In this webinar, we will delve into the theory and talk about the implementation in Pumas. We will demonstrate how this allows for faster and potentially more robust estimation of non-Gaussian models such as logistic regression, count data models, as well as ordinal regression models based on the FOCE approximation.
Check out the second webinar in the Pumas 2.0 feature series "FOCE Analysis of Discrete Data Models" for more information. Register now!