A reference library of data analysis models
12:00-14:30 (UK), 13:00-15:30 (Europe), 07:00-09:30 (US East coast), 21:00-23:30 (Australia East coast):
For many users, one of the barriers to modern data analysis tools is the need to have a background in coding. We have begun creating a common model language to make this easier across different platforms but we also think that the creation of a library of example models would be a useful starting point for users to build an appropriate model for their system. This session will therefore:
- Review the current status of the common model language
- Review relevant results of the ORSO survey
- Discuss how such a library could work in practice
- Canvas what types of example would be useful
The use of nested sampling in reflectometry
20:00-21:30 (UK), 21:00-22:30 (Europe), 15:00-16:30 (US East coast), 05:00-06:30 (Australia East coast):
Whilst Bayesian approaches are now common for estimating posterior distributions (to understand parameter uncertainty), data analyses that obtain the Bayesian evidence term for model selection (which of model A or B is a better way of treating the data?) are much scarcer. Nested sampling is a Bayesian statistical technique for estimating this evidence term, as well as being able to estimate the parameter posterior distribution.
- The first half of this session will have a brief introduction to nested sampling and its historical use in neutron reflectometry (45 minutes, speakers Andrew McCluskey and Andrew Caruana).
- The second half of the session will discuss possible future uses of nested sampling in reflectometry analysis (45 minutes).