Poisson Regression
A Poisson regression is another type of GLM. This requires integers or count data (i.e., 0, 1, 2, 3, …). For some situations, a Poisson regression can be more powerful (e.g., detecting statistically significantly trends) than a linear model or "Gaussian" regression.
During this exercise, we're going to build a linear regression using the lm() function and a Poisson regression using glm().
The objects x and y are loaded into R for you.
Den här övningen är en del av kursen
Hierarchical and Mixed Effects Models in R
Övningsinstruktioner
- Build an
lm()whereyis predicted byxand then print the summary. - Build a
glm()whereyis predicted byxwith a"poisson"distribution function and then print the summary to the terminal. - Examine the coefficient estimates for each and notice how only the
glm()produces statistically significant estimates forx.
Interaktiv övning med praktiskt arbete
Testa den här övningen genom att slutföra den här exempelkoden.
# Fit the linear model
summary(lm(___))
# Fit the generalized linear model
summary(glm(___, family = "___"))