I loved this brief video that out things i already know...into a different perspective. Thanks. I went to college in 1980's. I first came across Bayesian Statistics when I came across a paper on how spam filters work. But now I learnt that most of social science I learnt was frequentist. That's why they were descriptive. In that sense even inferential statistics was about the past. Things are very different when you use past data to predict the future.
After ages of scrolling through the internet to understand what probability means, I finally reach my destination. I always felt that the analogy of 'Probability of flipping a coin' has different interpretations, one being where we can confirm our beliefs after something has happened for a large frequency of time ( Now I can confirm that it is officially called as the FREQUENTIST APPROACH ) and other one where we just know that it is equally likely. Wrapping my head around these concepts and confirming my beliefs was really a painful one. I finally feel satisfied. Thanks Brian for this video.
Awesome video! I am currently taking a module called "Reliability and Risk Analysis" that evolves around probabilities, your video was clear and helped me learn a bit more about that subject
Very true. You could put a prior on the probability of them lying - ET Jaynes wrote some interesting stuff about this in the context of experiments to see if ESP exists.
Thank you, but one thing I still don't get about about the Bayesian way of thinking: if one is supposed to constantly update a belief via data, then where did the prior come from? How does one quantify a belief based on zero data?
If i tell tou that my coin is fair and we launch a coin. We toss it 99 times and it comes up heads, what is the probability that we get tails if you use both approaches?
Thanks! The Udemy courses are pretty different because they are structured from start to finish for a 4-10 hour course, whereas the RUclips videos are designed to be standalone. Currently, there are only very small areas of overlap.
There is only one flip of a coin, just as there is only one election. It makes sense for an F to say "given two sides, it's 50% probability the NEXT flip will be a head. It makes sense for a B to say "given the mood of the country, I'd wager even money the NEXT election will see X win." Our two punters, F and B, are using different forms and sources of data and related factors, but both are willing to wager at odds. That's the test of sincerity, willing to bet on risk vs. reward. Answer Darwin's question: is that ripe fruit worth the risk of a predator?
Years after university this finally makes sense to me, bless you sir!
I loved this brief video that out things i already know...into a different perspective. Thanks.
I went to college in 1980's. I first came across Bayesian Statistics when I came across a paper on how spam filters work. But now I learnt that most of social science I learnt was frequentist. That's why they were descriptive. In that sense even inferential statistics was about the past. Things are very different when you use past data to predict the future.
After ages of scrolling through the internet to understand what probability means, I finally reach my destination. I always felt that the analogy of 'Probability of flipping a coin' has different interpretations, one being where we can confirm our beliefs after something has happened for a large frequency of time ( Now I can confirm that it is officially called as the FREQUENTIST APPROACH ) and other one where we just know that it is equally likely. Wrapping my head around these concepts and confirming my beliefs was really a painful one. I finally feel satisfied.
Thanks Brian for this video.
Awesome video! This answers my fundamental question about the difference between the two, but as is with all things statistics raises 100s more.
Awesome video! I am currently taking a module called "Reliability and Risk Analysis" that evolves around probabilities, your video was clear and helped me learn a bit more about that subject
Thanks! This is very helpful
@2:57: The Bayesian here is very trusting of government sources.
Very true. You could put a prior on the probability of them lying - ET Jaynes wrote some interesting stuff about this in the context of experiments to see if ESP exists.
Thank you, but one thing I still don't get about about the Bayesian way of thinking: if one is supposed to constantly update a belief via data, then where did the prior come from? How does one quantify a belief based on zero data?
Finally understood what it means when we say intervals are random in frequentist paradigm.
The two systems are very well explained. Thanks
If i tell tou that my coin is fair and we launch a coin. We toss it 99 times and it comes up heads, what is the probability that we get tails if you use both approaches?
Probably very very small!
Thanks for the nice explanation. It may be a way to go before you get to the Bernstein- von Mises theorem, though. Still, good start.
AWESOME VIDEO
Was looking into your Udemy course. Love your teaching style. Are these lessons included in your Udemy course?
Thanks! The Udemy courses are pretty different because they are structured from start to finish for a 4-10 hour course, whereas the RUclips videos are designed to be standalone. Currently, there are only very small areas of overlap.
There is only one flip of a coin, just as there is only one election.
It makes sense for an F to say "given two sides, it's 50% probability the NEXT flip will be a head.
It makes sense for a B to say "given the mood of the country, I'd wager even money the NEXT election will see X win."
Our two punters, F and B, are using different forms and sources of data and related factors, but both are willing to wager at odds.
That's the test of sincerity, willing to bet on risk vs. reward.
Answer Darwin's question: is that ripe fruit worth the risk of a predator?
Amen brother.