Bayes probability

[majorendre]

Thank you, great video!

[magomah1]

thanx



[wangstick]

I have a question. Say you get tested 2 times. Intuitively, I would think the probability of, say, being positive and having it would increase substantially. But when I do the math by this same prob. tree method, I end up getting the same probabilities. What is the correct way of doing this? Or do the probabilities just not change no matter how many tests?

[1960sadm]

@wangstick

you need to start with 67/33 rather than 10/90 and follow same steps you increase the probability from 67% to 97.4% you have the disease if you test positive twice in a row!!.

[Khullah]

I just wish i had seen this before. I was over the formula for quite some time without understanding it. This is by far the best method to do it, and its how I did it. Thanks for posting, this will help a lot of people. For anyone who wants to go even further, I suggest to investigate on Bayesian Networks (things get really complex down there).

[nabecaydim]

you tested negative, what is the probability of you not having the disease: 0.914

[hicelina]

thank you

[mastersgta1]

Thank you so much for explaining this intuitively! Now I actually understand how to use it.

[Emjhey12]

great help! :D

[profribasmat217]

explanation is for a basic math college math class, think artists journalists etc., not a statistics course. Hence, terminology of statistics is not strictly employed, but rather the way a newspaper or lawyer would speak of them.

[sweetburlap]

It would be nicer if the terminology was a bit more strict. The term "accuracy" in this context is ambiguous at best (personally I would define accuracy as the proportion of true results ie. test result false in those without disease, test result true with disease). Sensitivity might have been a more appropriate term. (Although I do acknowledge that the terminology can get in the way of the basic principles)

It's a fantastic explanation of the principles involved nonetheless.

[shameelfaraz]

why did we multiply .92 and .08 by .10 ?

[danielearwicker]

@shameelfaraz Because it's 92% of the 10% that have it. So take 10% of the population, then take 92% of that small group. e.g. population of 1,000,000 people, take 10%, that's 100,000 who have the disease. Then of those, take 92%, which is 92,000 who get detected (population x .10 x .92) , which leaves 8000 who don't get detected (pop x .10 * .08).

[evelyn273]

Wonderful explanation! Thank you = )

[graw81]

Very good video! I also like to write out the extensive form to solve conditional probability problems. Thanks for posting.

[WyndiFosh]

Thank you so much! I get it. I have been so lost for so long. I appreciate you taking the time to upload this for all of us!

[kujta55]

Dear Sir,

Thank you a million times for explaining so simply. I have been going throw books and some other youtube videos and I still didn't understand the Bayes theorem.

Your video helped me so much, and now I fully understand it.

THANKS THANKS

[iym666]

Thank you---I wish my prof. at U of Chicago made things this simple!!!

[myutbest]

very nice; very innovative and precise. Excellent.

[myutbest]

very nice; with out talking about conditional explicitly; very innovative and precise. Kudos to you.

[Silveralex32]

nicely explained cheers

[isaiasperez]

Thanks!!! Very helpful!

[stoicblade69]

Wow, that helped a lot. Thanks.

[amin9846]

wunderful explanation

[thenetaficionado2]

Thanks for posting.

[twalsh90]

EXCELLENT!!

[jondesbrow]

Thank you, very clear and easy to understand, even for a novice such as myself.