still not sure I understand completely what you are saying, but the epistemological difference feels like this:
-
@odr_k4tana
still not sure I understand completely what you are saying, but the epistemological difference feels like this:goal Bayes - evaluate truth or falsity of theory, to do that make assumptions about diagnosticity of data
goal Freq - establish non-random nature of data pattern (phenomenon) and make assumptions about how that relates to theory
??
-
or in other words, from the perspective of the inherent symmetry of Bayes, where data follow from a hypothesis necessarily (ie by logical implication) there is no difference (conceptually or otherwise) between confirmation and falsification.
Most of the time, though, the data relate to the hypothesis less strictly (there is noise).
At that point Bayes and Freq have two different ways of dealing with that:
Bayes keeps the focus on a specific hypothesis, Freq focusses on the data1/2
-
2/2 as a result they make different choices on what to be specific about and what is more hand wavy....
-
which of these is more appropriate to a particular inductive problem then just depends on the context....(that is, the actual constraints available)
-
@UlrikeHahn looks like a reply but no @‘s made it this end
-
@andi ah!
computers...I was just so dizzy from my earlier triumph... -
Oliver D. Reithmaierreplied to Ulrike Hahn last edited by
@UlrikeHahn interesting. How would you categorise objective Bayesianism in this framework? The hypothesis part (i.e. specific prior) is subjective Bayes I presume.
-
Ulrike Hahnreplied to Oliver D. Reithmaier last edited by
@odr_k4tana in as much as it’s concerned with objective chances and their role in constraining rational belief I’d see it as part of working out circumstances when particular things are less hand wavy