Bayesian information criterion
Criterio de información bayesiano
- Date
- 1978 · confidence medium
- Place
- Jerusalem, Israel
- Credited to
- Gideon E. Schwarz
- Field
- Knowledge & symbols
What was discovered
A statistical criterion for comparing models that penalizes the number of parameters more heavily than the Akaike information criterion, derived by Gideon Schwarz from a Bayesian approximation to the likelihood.
What it led to
- Model selection in statistics and machine learning
- Systematic comparison with the Akaike information criterion
- Widespread use in genetics, economics, and ecology