In greta-dev/greta: Simple and Scalable Statistical Modelling in R

Hierarchical linear regression

A hierarchical, Bayesian linear regression model using the iris data, with random intercepts for each of the three species.

# linear model parameters
int <- normal(0, 10)
coef <- normal(0, 10)
sd <- cauchy(0, 3, truncation = c(0, Inf))

# hierarchical model for species effect; use the first species as the baseline
# like in lm()
species_sd <- lognormal(0, 1)
species_offset <- normal(0, species_sd, dim = 2)
species_effect <- rbind(0, species_offset)
species_id <- as.numeric(iris$Species)

# model
mu <- int + coef * iris$Sepal.Width + species_effect[species_id]
distribution(iris$Sepal.Length) <- normal(mu, sd)

greta-dev/greta documentation built on Dec. 21, 2024, 5:03 a.m.