Bayes Days 2000 at LANL
Sponsored by the Enhanced Reliability Methodology Project
of the Enhanced Surveillance Program
Minicourse on Bayesian Analysis in Physics
Prof. Volker Dose
Max Planck Institute for Plasma Physics
Garching bei München, Germany
Presented April 3-5, 2000 at the Los Alamos National Laboratory
Prof. Dose and his team at the Max Planck Institute have done seminal work
in exploiting the capabilities of Bayesian methods to solve very difficult physics
problems. In this lecture series, Dose summarizes many of those applications,
which are mainly in the area of plasma physics, but have much wider applicability.
High-quality videos of Prof. Dose's lectures are available at the LANL Research Library. Online streaming video versions can be viewed on the LANL network
LANL Media Theater.
As the image quality of this medium is fairly low, the viewgraphs below may be useful to have when viewing the video.
The viewgraphs for Dose's lectures are available in PDF (apologies for rotation of first viewgraph); each file is about 400 KB.
Some useful URLs on Bayesian inference
Bayesian principles (viewgraphs, PDF)
sum and product rule, marginalization, assigning probabilities
Cross section formulae for partial electron impact ionization of CH4 and H2 (viewgraphs, PDF)
parameter estimation, model comparison, posterior estimates
Outlier tolerant parameter estimation (viewgraphs, PDF)
pooling measurements of a physical constant with varying uncertainties, possibly incorrectly specified
Fitting scattered data (viewgraphs, PDF)
fitting models to data whose scatter greatly exceeds the quoted errors;
Thompson scattering in a Tokamak edge plasma
Deconvolution (viewgraphs, PDF)
exchange splitting of Nickel d-bands from inverse photoemission data;
Cu isotope abundancies from Rutherford backscattering spectra
Energy confinement in fusion devices (viewgraphs, PDF)
comparison of various models
Ken Hanson, DX-3, 667-1402, email@example.com
If you are interested in uncertainty analysis, participate in the
Information about last year's Bayes Dayes 1999.
Return to Ken Hanson's home page.