0
$\begingroup$

Given that bayesian probabilites are updated over time, how is this represented in a probability space? More specifically, how do we interpret the probability measure of such a space? That is, presumably such a probability measure is not preserved over time, whereas in frequentist statistics this is assumed to be invariant, so is a bayesian probability measure 'time-varying'?

$\endgroup$
3
  • $\begingroup$ In a frequentist world, if you take another sample, then you get other estimates, isn't that the same ? $\endgroup$
    – user83346
    Commented Oct 9, 2016 at 15:09
  • $\begingroup$ There really are no "Bayesian" probabilities, so I'm not sure what you mean. The term is just used to refer to a certain interpretation of probability, but the mathematics does not change in any way. $\endgroup$
    – dsaxton
    Commented Oct 9, 2016 at 15:35
  • $\begingroup$ You may be interested in checking more general thread: stats.stackexchange.com/questions/173056/… $\endgroup$
    – Tim
    Commented Oct 9, 2016 at 16:03

0