Intro material

• introductions
• course material: https://github.com/bbolker/quebec_2019 and https://bbolker.github.io/quebec_2019
• data sets
• please ask questions!

focus

• “advanced” end users (modelers, not statisticians)
• mechanistic rather than phenomenological models (cf. Breiman (2001))
• work with packages when possible …
• … but …

Rochester Digital Scholarship lab

• things used to be simple (lm, glm, nlme, lme4 …)
• now “let a hundred flowers bloom” (TMB, brms, greta, NIMBLE, …)

Dan Jeffries

outline

• likelihood estimation and inference
• nonlinear optimization methods
• simple MLE examples
• latent variables/random effects in nonlinear models
• (extended generalized linear mixed models)
• complex examples/nonlinear mixed models

tentative schedule

• 0830-1000: review of likelihood/inference, optimization methods, parameterization (from a geometric perspective)
• 1000-1030: break
• 1030-1200: simple examples with bbmle/optim
• 1200-1330: lunch
• 1330-1500: latent/mixed models
• 1500-1530: break
• 1530-1700: advanced examples with nlmer/TMB/etc.

references

Breiman, Leo. 2001. “Statistical Modeling: The Two Cultures.” Statistical Science 16 (3): 199–215. http://www.jstor.org/stable/2676681.