Introduction
What this guide covers
In the biomedical and health sciences, response variables of clinical interest rarely satisfy the assumptions of normality and homoscedasticity required by the classical linear model estimated via ordinary least squares (OLS). Treatment success or failure (binary), bacterial colony counts or seizure episodes (counts), hospital cost or time to an event represent typical examples.
Overview
Topics covered
- Introduction
- Regression Models for Non-Normal Response Variables
- Limits of the classical linear model in biomedicine
- The Exponential Dispersion Family
- Architecture of Generalized Linear Models (GLM)
- 1. The random component
Related concepts
Main connections
Compatible notation
Mathematical sample in LaTeX
The expression is rendered accessibly, with its LaTeX source preserved below.
View LaTeX source
Y_i = \mathbf{x}_i^T \boldsymbol{\beta} + \varepsilon_i, \qquad \varepsilon_i \stackrel{\text{i.i.d.}}{\sim} \mathcal{N}(0, \sigma^2)