Goodness of fit

From Open Risk Manual
The printable version is no longer supported and may have rendering errors. Please update your browser bookmarks and please use the default browser print function instead.

Definition

The Goodness of fit of a statistical Risk Model describes how well it fits a set of observations, where fitness is typically a quantitative measure of discrepancy between observed (realized values) and the values expected (predicted, forecast) by the model.

There is a wide range of possible measures (See wikipedia collection). They can be used both to guide the process of Model Development and assist with the quantitative aspects of Model Validation