I usually never comment maths videos, but tbh this is litteraly the first time i see only one person who explain a concept properly, so thank you for that
Before this I was going through other TH-cam videos. Everyone saying missing data. But in the first min you said it's missing col of a sheet and not row. This cleared my confusion. Thank you sir
As a biostats PhD student, you made this click for me in 30 minutes while three separate classes that covered the theory behind this algorithm weren't able to do that!
I understand that in the coins example in the beginning, you used the binomial distribution to get the likelihood. Is it true that you have to replace that with the log likelihood ?
The whole video is a coin flip example, so I’m not sure what you mean. We don’t use log likelihoods in the chart where we solve things, but if you watch the whole video we derive the method for the chart by using log likelihoods. Likelihoods and log likelihoods are the same thing, one just has a log.
Great vide, but I have a quick question, I did not exactly get where the missing data is? You already showed the 10 coin flips that we do not know which coin each came from. As you stated at the beginning, shouldn't we have an entire column missing? Shouldn't the data be induced somehow from somewhere? So if I have observations for the columns, say, A, B and C, and I'm missing observations for columns D and E (entire columns are empty). How would the initial data be induced or imputed in the first place so we are able to work with it? Am I missing something?
The missing column in the data is whether the flips are from coin A or B. The initial guesses for the missing column are random and we will often run the algorithm multiple times with different starting guesses so that we don’t get stuck in a local maximum.
This has nothing to do with the video, but I tried to buy a biased coin from Amazon - no result -, and searching for other means to get one (like Galton Boards, they would be great in class), I had to find out that biased coins probably do not exist. Help me out if you know more about that.
I believe you could make a biased coin by welding a straight wire with specific length to one side that is sticking straight outward from the middle. This way when flipping the coin, the wire stick will operate as a weight trying to balance the coin pointing downward, like a keel for a ship, so more times will the coin point with the wire downward. I've not tried this but I believe this could work.
I usually never comment maths videos, but tbh this is litteraly the first time i see only one person who explain a concept properly, so thank you for that
I'm feeling a Statquest vibe here and this is very good! Keep the good work, bro! Subscribed!
Thank you!
Before this I was going through other TH-cam videos. Everyone saying missing data. But in the first min you said it's missing col of a sheet and not row. This cleared my confusion. Thank you sir
I totally agree, I used to be confused by the same thing :)
most simple yet in depth analysis found so far, very intuitive..thanks a lot, because of you all my doubts got cleared
you gonna blow up soon! keep up the good work
Thank you! :)
Thanks!
Thank you!
As a biostats PhD student, you made this click for me in 30 minutes while three separate classes that covered the theory behind this algorithm weren't able to do that!
I always felt the same way until I tried making a video about it. :) So happy to hear it helped you and that somebody watched the whole thing!
Subscribed. Very very very good content. appreciate it
Thank you!
I understand that in the coins example in the beginning, you used the binomial distribution to get the likelihood. Is it true that you have to replace that with the log likelihood ?
The whole video is a coin flip example, so I’m not sure what you mean. We don’t use log likelihoods in the chart where we solve things, but if you watch the whole video we derive the method for the chart by using log likelihoods. Likelihoods and log likelihoods are the same thing, one just has a log.
nicely explained :)
Great vide, but I have a quick question, I did not exactly get where the missing data is? You already showed the 10 coin flips that we do not know which coin each came from. As you stated at the beginning, shouldn't we have an entire column missing? Shouldn't the data be induced somehow from somewhere? So if I have observations for the columns, say, A, B and C, and I'm missing observations for columns D and E (entire columns are empty). How would the initial data be induced or imputed in the first place so we are able to work with it? Am I missing something?
The missing column in the data is whether the flips are from coin A or B. The initial guesses for the missing column are random and we will often run the algorithm multiple times with different starting guesses so that we don’t get stuck in a local maximum.
@@statswithbrian Great thanks for the clarification and the great video!
Thank You
This has nothing to do with the video, but I tried to buy a biased coin from Amazon - no result -, and searching for other means to get one (like Galton Boards, they would be great in class), I had to find out that biased coins probably do not exist. Help me out if you know more about that.
I honestly don’t know if it’s physically possible but I’m guessing not. Would love an update if you find one.
I believe you could make a biased coin by welding a straight wire with specific length to one side that is sticking straight outward from the middle. This way when flipping the coin, the wire stick will operate as a weight trying to balance the coin pointing downward, like a keel for a ship, so more times will the coin point with the wire downward. I've not tried this but I believe this could work.