3 ms·
For Bayesian Inference, the example in the Wikipedia article is good (who doesn't like cookies?): https://en.m.wikipedia.org/wiki/Bayesian_inference#Examples ht
by dwattttt 2y ago
For Bayesian Inference, the example in the Wikipedia article is good (who doesn't like cookies?): https://en.m.wikipedia.org/wiki/Bayesian_inference#Examples https://en.m.wikipedia.org/wiki/Bayesian_inference#Examples
When gathering more evidence, you'd use your new belief about which cookie bowl Fred has as P(H1)=0.6 and P(H2)=0.4
- eru 2y agoThat just explains Bayesian inference in general, which is useful, but not what I'm after. I was specifically interested in the application to binary search / bisection in the presence of flaky tests.
- dwattttt 2y agoYou apply the same process, but your hypotheses are "bug is before/after this bisection point". Your "probability of evidence given hypothesis before/after" are where you incorporate your guess about the tests flakiness. Still works even if you don't have "true" numbers for the tests flakiness, just won't converge as quickly