A generalized linear model (GLM) is a statistical framework that extends standard linear regression to handle outcomes that don't follow a normal distribution, including yes/no outcomes, event counts, and skewed distributions.
Every GLM is built on three components.
- The random component, which defines the probability distribution of the response variable;
- The linear predictor, which combines the input variables; and
- The link function, which connects the two and keeps predictions within valid bounds.
Common types of GLMs include logistic regression for binary outcomes, Poisson regression for count data, and gamma regression for continuous values that are strictly positive and skewed.
Share