ABC for parameter inference and model selection in dynamical systems (2)
As it happens, I did not get the final and published version of the paper “Approximate Bayesian Computation scheme for parameter inference and model selection in dynamical systems” by Toni, Welch, Strelkowa, Ipsen and Stumpf, as of June 12, but an earlier version of April 31, 2008, where there was a typo. In the final version, the difficulty I had yesterday has vanished in that the importance weight used in the ABC SMC algorithm is now correct (and corresponds to the ABC PMC weight as well). There are a few minor points of contention remaining with the appendix A of the paper, in particular in the definition of ηt around (A3) but the algorithm stands corrected. Sorry if this created any confusion!
Ps-I did not even know the Journal of the Royal Society Interface existed, as it is not an usual outlet for Statistics papers… It appears to be an equivalent to the US PNAS (and in a more remote manner to the French Notes aux Comptes Rendus de l’Académie des Sciences). To end up on a light note, the current news heading on the Royal Society webpage is “Fluorescent mice help shed light on the spread of a dangerous disease“. Of course! The more fluorescent, the more light!