News for hackers. Stuff that matters.
Re:Frequentists should more often consider using Bayesian me... by kobeya on Saturday November 19, 2016 @04:43AM (#12992721)
The one that yields the simpler solution.
Re:Frequentists should more often consider using Bayesian me... by murbard2 on Saturday November 19, 2016 @03:28PM (#12995442)
How do you know the solution is any good? I'm not going to go all Socrates on you so I'll jump to my point: when discussing bayesian statistics, we're touching something very profound about epistemology that can't be swept under the rug in the name of pragmatism. We're dealing with the core philosophical underpinnings of what it means to "know" something.
Re:Frequentists should more often consider using Bayesian me... by dragandj on Saturday November 19, 2016 @05:02AM (#12992751)
How do you know which one leads to any good solution, let alone a simple solution?
Re:Frequentists should more often consider using Bayesian me... by kgwgk on Saturday November 19, 2016 @05:10AM (#12992773)
Another option is to choose the one that yields the (more) correct solution.