2019/02/18 by Gilles Kratzer, Reinhard Furrer, Kratzer, Gilles +1
Computer Science · Environmental Science · Health Professions · #Bayesian Modeling and Causal Inference #Health, Environment, Cognitive Aging #Artificial Intelligence in Healthcare
paper · pdf · doi:10.48550/arxiv.1902.06641
Bayesian network (BN) modelling is extensively used in systems epidemiology.\nUsually it consists in selecting and reporting the best-fitting structure\nconditional to the data. A major practical concern is avoiding overfitting, on\naccount of its extreme flexibility and its modelling richness. Many approaches\nhave been proposed to control for overfitting. Unfortunately, they essentially\nall rely on very crude decisions that result in too simplistic approaches for\nsuch complex systems. In practice, with limited data sampled from complex\nsystem, this approach seems too simplistic. An alternative would be to use the\nMonte Carlo Markov chain model choice (MC3) over the network to learn the\nlandscape of reasonably supported networks, and then to present all possible\narcs with their MCMC support. This paper presents an R implementation, called\nmcmcabn, of a flexible structural MC3 that is accessible to non-specialists.\n