Bayesian posterior probabilities: revisited - Quantitative Biology > Populations and EvolutionReport as inadecuate




Bayesian posterior probabilities: revisited - Quantitative Biology > Populations and Evolution - Download this document for free, or read online. Document in PDF available to download.

Abstract: Huelsenbeck and Rannala 2004, Systematic Biology 53, 904-913 presented aseries of simulations in order to assess the extent to which the bayesianposterior probabilities associated with phylogenetic trees represent thestandard frequentist statistical interpretation. They concluded that when theanalysis model matches the generating model then the bayesian posteriorprobabilities are correct, but that the probabilities are much too large whenthe model is under-specified and slightly too small when the model isover-specified. Here, I take issue with the first conclusion, and insteadcontend that their simulation data show that the posterior probabilities arestill slightly too large even when the models match. Furthermore, I suggestthat the data show that the degree of this over-estimation increases as thesequence length increases, and that it might increase as model complexityincreases. I also provide some comments on the authors- conclusions concerningwhether bootstrap proportions over- or under-estimate the true probabilities.



Author: David A. Morrison

Source: https://arxiv.org/







Related documents