Generalized linear model

The observed variable might not always be amenable to the assumptions of normal distribution and, in those cases, using a linear model is not a good idea. Instead, we can use a GLM, which allows the observed variable to have an error distribution different to the normal distribution.

Modeling count data using Poisson GLM

Suppose we have the migration rates for each of the areas we have considered so far. This is the count data, and if it can be assumed that the migration rate is constant, we can use the Poisson model to get the probability of a specified rate of migration. In a Poisson model, the mean and variance are the same.

Using the glm() function, we can fit a different GLM model. Here we model...