Random Variable Representation
From Open Risk Manual
Definition
Random Variable Representation is the computational representation of a Random Variable concept, e.g. in a form that can be used in digital computers.
Representation Forms
Depending on its nature, a random variable may
- explicity specified through some functional form (see. e.g. Risk Distribution)
- defined (implicitly) as the compound Sampling outcome of a complex system (realized in practices via Simulation Models), in which case the representation is only as accurate as the Sampling quality
- specified as a Histogram that is either exact (for discrete random variables) or constructed as a sample
Issues and Challenges
- Mathematically random variables may be continuous or have continuous support. In computer representations some discrete approximation must be applied
- Mathematically random variables may have a range extending to infinity. In computer representations some finite approximation must be used as a proxy
- Representing random variables via a sample may be materially inadequate to capture low-likelihood realizations