Estimating the number of undetected errors: Bayesian model selection | IEEE Conference Publication | IEEE Xplore