Bayesian inference and computation
The defining feature of the Bayesian approach to statistical inference is that uncertainty is expressed through probability.
The defining feature of the Bayesian approach to statistical inference is that uncertainty is expressed through probability.
The mechanism for this is Bayes theorem, which updates prior probability distributions for unknown parameters of a statistical model to posterior distributions, when data are observed. The Bayesian approach also provides a coherent mechanism for prediction in complex applications.
Research in Bayesian inference in S3RI focuses on enabling users to apply the methodology in complex applications. Much of the research within the modelling theme at large involves developing Bayesian methodology. Prominent examples are described under each of Complex categorical data analysis, Graphical models, Modelling processes in space and time, Statistical methods for missing data, Statistical disclosure risk assessment and control and Applications in the social sciences. Related research into Bayesian approaches to design of experiments is described under the Design of Experiments theme.
| Related Projects | Status | Type |
|---|---|---|
| Novel statistical approaches to chemical, biological or radiological source term estimation | Active | Other |
| Estimation of structural dynamic parameters at higher | Active | Other |
| Bayesian model determination | Active | Other |