Prior And Posterior - Intro to Statistics

In this video, Udacity founder & AI/ML (artificial intelligence & machine learning) pioneer, Sebastian Thrun, gives you a comprehensive introduction to statistics, prior and posterior.

---

Connect with us on social! 🌐

[yonashlomo]

Starting :049 there's a very confusing mistake in the video: Instead of saying that Posterior P(C|Pos) was what was written -- it is actually P(C and Pos)
Correcting this video would be helpful in reducing confusion.

[xsli]

The author Sebastian Thrun is really a famous Stanford professor. He has mentioned that the formula is incomplete. Both formulas at the end of the clip[ ) need to be divided by a normalizing constant, which is P(C)xP(Pos|C) + P(~C)xP(Pos|~C).

[ThachDo]

Your posterior formulas are wrong in term of Bayer theorem, P(Pos) should be multiplied on the left-hand side of the equations.

[AmithAdiraju]

Those of you who're wondering, there's nothing wrong with the formula, the video is just "INCOMPLETE", I've taken their full Stat Course, in the next part of the video they divide the normalizing constant to both posteriors, that is by Probability(Positive Test) ....

And when you multiply P(C) * P(Pos/C), you get P(C, Pos) which is not same as posterior, i.e., P(C/Pos) ; to get P(C/Pos) we need P(C, Pos) / P(Pos), this is same as P(C) * P(Pos/C) / P(Pos) ; for brevity purposes the author has written in a small shortcut formula

Before jumping to conclusions quickly, please understand that, they've put this video just as a "Bite-Sized Trailer" to the course, it's not meant to be a full fledged intro to Bayes theorem, it's Udacity's style to break-up videos into really small chunks ( 4 seconds long even ), they follow everything up in a really small sequences ... It's sere naivety to think a platform like udacity would make such a boo-boo mistake ... Folks, stop thinking that you're a genius for a second and try to reason before judging something

[maoxuliu]

He made a mistake which was confusing us. The formula should be P(Pos, Cancer) = P(Pos | Cancer) * P(Cancer)

[jakob]

Nice introduction but the rest of the video gets a little confusing when g is written instead of 9

[mikiallen]

can we say that prior probability is the same as "unconditional probability" i.e. in the absence of any explanatory variable, according to your example, p(c) = 0.3, i, e, probability of getting a cancer is 30% in absence of all explanatory variables / risk factors as used within "epidemiology", and if this is not the case, how one should differentiate between the two if that even makes sense ?
merci

[keyleegibbons]

Why when professors talk about Bayes Theorum do they always use the cancer example. I wish there was a less depressing example. Give me the probability that I get to have puppies given that my dog is fat. Also, it's wrong.

[igorcryptor]

It's really seems that this video is a part of bigger context, that is not explained here...

[tosifkhan]

please do not watch this incomplete video

[jackpainting]

I wouldn't call this an 'intro' to statistics.

[Shamssali]

this video has been here for 11 years delete this already

[zaher]

the way mr sebastian illustrate is really so bad

[jackthederp]

can't understand what you are saying, might want to speak your native language next time because I don't think any English speaking people can understand you.