CoRe in CiRM [6]

This is the last day at CiRM for our “research in pair” working on the new edition of Bayesian Core. I mostly completed the revision of the chapter on Bayesian imaging, including stuff on Markov random fields and mainly on ABC algorithms. All that remains to be done about this chapter is the processing of the Scottish lip cancer dataset from an MRF perspective. I included the comparison below of the ABC approximation (variability) with the truth in the case of the normal mean test.  Jean-Michel has almost finished the chapter on hierarchical modelling, before I get a go at the style of the chapter. Thus ends a very productive ten-days where we dealt with five chapters out of nine, the remaining four being less prone to deep modification in this new edition…