Department of Biostatistics Seminar/Workshop Series
Combining computational models with physical observations for inference and prediction
Dave Higdon, PhD
Los Alamos National Laboratory
Wednesday, October 24, 1:30-2:30pm, MRBIII Room 1220
Combining physical measurements with computational models is key to many scientific investigations. In order to use computational models to help in the inference process, one must typically estimate model parameters and account for the difference between model and reality. This talk briefly surveys some of the many approaches taken in this field. The approach taken typically depends on specific features regarding the problem at hand - data size, amount of empiricisms in the model, availability of data, extent of extrapolation required, etc. This talk will consider an example from cosmology (and possibly radiative shock physics) to highlight how one might constrain unknown model parameters and estimate prediction uncertainty.