Fermat number inference - 2/2 Re:Stanford Challenge

[dmlaurent]

Well this is a change to music videos!! I enjoyed the presentation and many more are needed like it. Please continue these lectures. Bayesian probably is extremely important concept.

[clray123]

While P(f(0) prime | M1) is quite clearly 1, I don't see how you come up with P(f(0) prime | M2) being also 1? Or P(f(0),f(1) prime | M2) being 0.75?

[trondreitan]

I described this in clip 1, though I may have been a little unclear. In Model 2, the sequence so far is prime by pure chance. As the prime numbers are more sparsely distributed the farther you go, the probability of an arbitrary number being prime is estimated as the rate of prime number up to the number in question, starting at 2. Pr(f(0) prime|M2)=2/(3-1)=1. Pr(f(1) prime|M2)=3/(5-1)=75%.

[scooterboy94]

what the fuck is that gay peice of shit how is that a tackle turning wrong you fat fuck i hope you fucking die of being gay

[StarF68]

Wow, you're an idiot!

[VirtualDscourse]

Interesting.

[websnarf]

Well it is interesting me, and I'm glad you did them! But I have mathematical bent. I'm actually curious if you had some references to these Bayesean formulas so that I can go study them more thoroughly myself.

[trondreitan]

Thanks! I've posted a couple of wikipedia links at the end of the video description. From a glance, the articles seem to make sense. I can also recommend Jaynes: Probability theory - the logic of science. I can not recommend Berger: statistical decision theory and Bayesian analysis, but that was my intro do the field.