Generalized linear model
Modelo lineal generalizado
- Date
- 1972 · confidence high
- Place
- Harpenden, England, United Kingdom
- Credited to
- John Nelder y Robert Wedderburn
- Field
- Knowledge & symbols
What it is
A statistical framework generalizing ordinary linear regression by relating the response variable to the predictors through a link function and allowing the error distribution to vary, unifying linear, logistic and Poisson regression; formulated by John Nelder and Robert Wedderburn.
What it led to
- Unification of linear, logistic and Poisson regression
- Standard statistical software (GLM in R, SAS, etc.)
- Basis for generalized linear mixed models and additive models