Modeling
Overview
Probabilistic analysis requires three essential elements including: (a) predictive models that relate the process outcome to a set of observable process variables, and a set of unobservable model parameters, b) statistical distribution models that describe the behavior of observable process variables, and c) statistical distribution models that define the uncertainty of the unobservable predictive model parameters.
This technology area focusses on development of modeling techniques to:
- Construct the empirical predictive models, the distribution models for observable process variables based on available data size, and the uncertainty models for the unobservable predictive model parameters;
- Develop model verification and validation techniques to ensure reliability of the models; and
- Assess, update, and calibrate the models when new data becomes available.