We should start unconventional models with the distribution that has the biggest entropy b/c it spreads probability most evenly while remaining true to the info about the parameters and outcomes that we know.

This chapter introduces you to GLM and maximum entropy.

The chapters that follow build computational skills for working with different flavors of GLM. Ch. 10 addresses models for count variables Ch. 11 addresses models for ordinal variables and mixtures Ch. 12 makes use of binomial count models (and relies on Ch 10) Ch. 12 and onwards deals with multi-level models.



joepowers16/rethinking documentation built on June 2, 2019, 6:52 p.m.