https://en.wikipedia.org/wiki/Metropolis%E2%80%93Hastings_algorithm Just curious... And what do Markov Chains have to do with it?
Oh! Yeah!... [video=youtube;Ws63I3F7Moc]https://www.youtube.com/watch?v=Ws63I3F7Moc[/video] Markov... Hmmmm.... [video=youtube;x_jRQBGKPaA]https://www.youtube.com/watch?v=x_jRQBGKPaA[/video]
What do you want to know? Metropolis-Hastings finds the maximum of a function by a kind of random walk. There is some probability of going in the unintuitive direction (away from a maximum point) so that one doesn't get stuck in local minima. The random walk is from a Markov Chain. Markov Chain Monte Carlo is used because it's relatively quick compared to finding the slope and following that (at least in problems with many dimensions).
So if one has a proper narrative stick, one can gets back to work???... [video=youtube;K-bJnxJl4ZA]https://www.youtube.com/watch?v=K-bJnxJl4ZA[/video]
Is there a Cambridge comma? [video=youtube;8Xjr2hnOHiM]https://www.youtube.com/watch?v=8Xjr2hnOHiM[/video]