Difference between revisions of "Model Usage"
From Open Risk Manual
Wiki admin (talk | contribs) |
Wiki admin (talk | contribs) |
||
Line 1: | Line 1: | ||
== Definition == | == Definition == | ||
− | '''Model Usage''' denotes the precise manner in which [[Model Outputs]] are translated into business actions. | + | '''Model Usage''' denotes the precise manner in which [[Model Outputs]] are translated into business actions. It is the context into a certian model is put to use. |
− | Usage can be either in the form of automated / unsupervised action or instead produce signals / metrics which are to be evaluated by humans before any action is taken | + | == Usage Modes == |
+ | Usage can be either in the form of | ||
+ | * automated / unsupervised action or instead | ||
+ | * produce signals / metrics which are to be evaluated by humans before any action is taken | ||
== Examples == | == Examples == |
Latest revision as of 19:43, 11 March 2024
Definition
Model Usage denotes the precise manner in which Model Outputs are translated into business actions. It is the context into a certian model is put to use.
Usage Modes
Usage can be either in the form of
- automated / unsupervised action or instead
- produce signals / metrics which are to be evaluated by humans before any action is taken
Examples
Mode usage can take many distinct forms:
- yes/no decisions in accepting or proposing a risk/transaction
- direct input into pricing decisions
- direct input into hedging decisions
- input into setting of limits
- internal and / or regulatory capital calculation
Model usage includes the use of Overrides, re-calibrations or any other model adjustment mechanisms that may intermediate and modify the model output before final action.
Issues and Challenges
- Comprehensive documentation of all important uses of a model can be lacking in Model Documentation because once models are in place users may not follow strict Model Governance