4 ms·
With bayesian probability a single observation can be informative (in the sense that a single observation leads to a change in distribution over hypotheses or p
by Vetch 5y ago
With bayesian probability a single observation can be informative (in the sense that a single observation leads to a change in distribution over hypotheses or parameters). It's a matter of narrowing down beliefs on weighted possibilities though, not a definitive answer on the state of reality.
- GoblinSlayer 5y agoThe author doesn't rely on observation. They propose combinatorial models, obtain large time estimates and notice that observation is consistent with these estimates because sample size one, it's a coincidence.