Originally Posted by
klar
uhhh. I don't think I understand what you mean. Bayesian problem as in a logical cause and effect issue? "Statistical" is perhaps the wrong word for the methods in question, maybe "empirical" would be better?
If X% of skier triggered avalanches occur in steep slopes with a northerly aspect, avoiding steep slopes with a northerly aspect will decrease your chances of triggering an avalanche to Y%, compared to a chance of Z% (Z>Y) when you select a slope at random from slopes of all aspects and angles. What do you mean by historical rate in this context?
Should the posterior distribution be reevaluated as more information becomes available (perhaps to incorporate additional "causal" variables), with the result then serving as the basis for a new model? yes.
Bookmarks