plom-pipe: Bayesian inference for compartmental models, with a Unix flavour
- AI4TS

The main motivation behind the open source library plom-pipe is to reduce the technical friction that prevents modellers from sharing their work, quickly iterating in crisis situations, and making their work directly usable by public authorities to serve decision-making. An illustration of this problem is the fact that the scientific production on the 2009 H1N1 epidemic has peaked three years after, in 2012, and that the vast majority of this production takes the form of static pdf files. Even when the corresponding code is shared, it takes a substantial amount of time and efforts for any other person than the author to at least reproduce the published results. This document gathers the methodological aspects on which the plom-pipe library is built on.
View on arXiv