you explain this very well. only if my teacher would have done the same thing, so i wouldnt have to teach myself via youtube. if im not learning stuff in class but i am learning here, the teacher is to be blamed. and why am i paying so much in tuition for some incompetent teacher at university?
WateryIce54321 true. and thats why often after each lecture i go online to have a different teacher explain the same concept, and then i totally get it.
+sorooshusa You know funny thing, I actually met someone who actually thinks that there is no such thing as a good or a bad teacher. So, apparently, it's all up to the student to learn the material. So, according to him, the teacher is not to be blamed but you. It sounds like bs though.
+TheKiosk94 at that point they're saying it doesn't matter who is teaching because you're going to teach yourself. which, is true, but sad. kind of begs the question, what do you think is the point of having a teacher?
As I am also trying to teach myself through RUclips, I come across comments like this all the time. I think slowly universities are realizing that just having a Ph.D. doesn't mean you know how to teach anything. In my opinion, all university instructors should be required to undergo basic pedagogical training before teaching any courses. It's sad how many of my colleagues with Ph.Ds have no clue about how the brain works, or how memories or formed, or just more generally, how learning works.
Thank you for the explanation. These are mostly just bookmarks that I am using for myself as I am preparing for a final exam in a month and really want to make sure that I have all of the resources needed when I am to hardcore study. 8:38 If MGF is known, then in some cases, computing expected value/second moment/third moment can become easier than using the standard definition of expected value. Do this by deriving and evaluating with t = 0. 11:04 Can use taylor series expansion to represent the MGF 14:30 Example: Find the expected value of Y = g(x) when the MGF of x is given.
Great videos. Clearly explained. I have subscribed to your channel. Please make videos of hypothesis testing, estimation theory, multivariate analysis, time series analysis, analysis of sample surveys & stochastic processes. Thank you.
Thank you so much for your videos you are so helpful and watching gets me excited about stats. I would jump for joy if you came out with some videos with practice problems
certains parts of this vid makes so much more sense, and then your video helped click some things looking at my prof notes of a review practice problems
It didn't, but it didn't need to. That was just setting up the question. It may as well have said; "after some period of time, let's call it 5 years..." or whatever. The actual mathematics question in there, if you want to call it that, is just that you have a random variable x, with the given MGF and you needed to calculate the expected value of g(x), where g(x) = 100(0.5)^x. That three years was just a random time, and was to set up the question and for it to have a tie to some real world situation. If you think about it, it wouldn't make sense for the value of the pice of equipment to be random at the present time... as they would know what it was worth then. Does that help?
What would you then do with that E(X)? Notice we want the E(Y), which is the function we were given at the beginning, we can't just simply sub in that E(X)... Get it?
Really good video.. couple of queries, though.. a) what is the significance of moment, physical interpretation b) from where moment generating function originate? c) what if t assumes other numerical values? does it have any significance.. I would be glad if these are answered... TIA
Great questions! Moment generating functions are special cases of characteristic functions, which are Fourier transformations of probability density functions. I may need to make another video just to answer your thoughtful questions. Your questions also lead me to believe you are capable of getting some answers if I point you to a couple of resources: From a historical perspective, the ideas of MGF were present in the book “Docrtine of Chances” by Abraham de Moivre (1667-1754). Mackey, George W. Harmonic analysis as the exploitation of symmetry-a historical survey. Bull. Amer. Math. Soc. 3 (1980), no. 1, part 1, 543-698 gives a great historical background of harmonic analysis and Fourier Transforms. For an intuitive and visual description of Fourier transforms, I recommend a video by 3blue1brown: ruclips.net/video/spUNpyF58BY/видео.html. It explains that Fourier transforms decompose a signal into its components. It also does a good job of connecting to the idea of detecting the center of mass (c.f., the expected value can intuitively be thought as the center of mass of the probability density function).
Good video! I have one question about your comment on calculating E(x). Instead of using the MGF, is it possible to calculate E(x) directly. If possible, show me the detail please. You just said that it is just difficult in the video. Thank you in advance.
The mgf for a Uniform Discrete is not defined when t=0. I can still find moments by taking limit when t-->0 of the various derivatives. Is it actually incorrect when we say that the first moment is the first derivative evaluated at 0. Is it more correct to say that the first moment is the limit as t-->0 even if the mgf is not defined at t=0.
M'(x)=\frac{d}{dt}\left[ \frac{3}{4}(1-\frac{e^t}{4})^{-1} ight] =\frac{-3}{4}(1-\frac{e^t}{4})^{-2}\frac{-e^t}{4} for t=0 M'(x)=\frac{-3}{4}(1-\frac{1}{4})^{-2}\frac{-1}{4} =\frac{1}{3} So wolfram was wrong and Actuarial Path is correct =D
I am taking probability lecture at my school and professor gave an assignment but I couldn't solve 1 question from it. question askes that if X is a binomial random variable with parameter (n,p) and Y=X^2 . find the variance of Y by using characteristic function of X. after watching this video I would solve this question in that way; Var(Y) = E(Y^2) - [E(Y)]^2 Var(Y) = E[x^2)^2] - [E(x^2)]^2 after that point I thought that I need to take fourth derivative of the characteristic function of X and take second derivstive of characteristic function of X and take its square root am I right?? characteristic function of X = Φx(w) = (1 - p + pe^(jw))^n I am not sure if this thought is correct can you help me?
thank you for the great video! but the mic is just too much of a treble. With volume set to max at the beginning of video my ears pained and then I jumped from the seat. Could you please fix and replace the audio if you have time. The crystal voice was a bit of a pain for my ears
Hello. Thank you for your amazing explanation. I just have one question, when we do series expansion for Mx(t), how come it is E(tx + tx^1)? If the function we're considering is e^(tx), how come it's not e^x and e^2x? Where does the exponential go? Thank you :)
hope those that disliked this video are not human. This is one of the best tutorial videos i have ever seen in my life.
1 hour before my exam and this man is always here to save my ass
you explain this very well. only if my teacher would have done the same thing, so i wouldnt have to teach myself via youtube. if im not learning stuff in class but i am learning here, the teacher is to be blamed. and why am i paying so much in tuition for some incompetent teacher at university?
+sorooshusa Because explaining for a general audience is a gift very few have :(
WateryIce54321 true. and thats why often after each lecture i go online to have a different teacher explain the same concept, and then i totally get it.
+sorooshusa You know funny thing, I actually met someone who actually thinks that there is no such thing as a good or a bad teacher. So, apparently, it's all up to the student to learn the material. So, according to him, the teacher is not to be blamed but you. It sounds like bs though.
+TheKiosk94 at that point they're saying it doesn't matter who is teaching because you're going to teach yourself. which, is true, but sad. kind of begs the question, what do you think is the point of having a teacher?
As I am also trying to teach myself through RUclips, I come across comments like this all the time. I think slowly universities are realizing that just having a Ph.D. doesn't mean you know how to teach anything. In my opinion, all university instructors should be required to undergo basic pedagogical training before teaching any courses. It's sad how many of my colleagues with Ph.Ds have no clue about how the brain works, or how memories or formed, or just more generally, how learning works.
Thank you for the explanation. These are mostly just bookmarks that I am using for myself as I am preparing for a final exam in a month and really want to make sure that I have all of the resources needed when I am to hardcore study.
8:38 If MGF is known, then in some cases, computing expected value/second moment/third moment can become easier than using the standard definition of expected value. Do this by deriving and evaluating with t = 0.
11:04 Can use taylor series expansion to represent the MGF
14:30 Example: Find the expected value of Y = g(x) when the MGF of x is given.
ive got my final in 8 hrs (gotta sleep somewhere in there too), and it's not looking good lol but this video is my last hope
thank you for all the vids, going in a month to take the exam p. These vids are for sure a great refresher
The way you have you change the MGF of X to MGF of Y is very Neat! I was confused at first as to how it would be done.
Fantastic video! much better for learning and understanding than my actual lectures!
This guy is damn good explaining topics my professor of 30 years cannot do
ok.... this was very helpfull... thanks sir... love from india
Great videos. Clearly explained. I have subscribed to your channel. Please make videos of hypothesis testing, estimation theory, multivariate analysis, time series analysis, analysis of sample surveys & stochastic processes. Thank you.
Thank you so much for your videos you are so helpful and watching gets me excited about stats. I would jump for joy if you came out with some videos with practice problems
This is great!!! I'll recommend it to my fellow comrades.
thankyou so much the lecture is astounding...
certains parts of this vid makes so much more sense, and then your video helped click some things looking at my prof notes of a review practice problems
Just a quick reminder it is |r|
I was thinking the same thing, And I thought he would fix it when he kept saying neighborhood of 0.
Your explanation is very good. I watch your videos.
Thanks a lot
Why didn't we have to take the derivative of the MGF before plugging in ln(.5)?
Thanks! I got some basic understandings of this topic in my head.
Thanks Sir.... may God bless you and your family
Now is 2020. This really helps me in my Probability Theory. Thanks
Thank you so much. It helps A LOT!
Very well explained. Thanks a lot to make the concept clear
Very simple explanation 👌👍
Love from India ❤️❤️
good vid
srsly fuk mgfs tho
ungrounded data ke leya relation between raw and central moment
It's important bro! Lol.
Good work sir
Explained very nicely!! Thank you very much!!!
Now ,it's very clear to me.Thank you sir..
Thank you for this video. Very clear and helpful!
I had several "a-ha" moments watching this (n.p.i.). tyvm
Amazing vid! Keep up the good job, mate!
Thanks alot, I have clearly gotten the concept behind m.g.f ,thanks alot once again
much clear now after watched this video! thanks!
How would you change the problem if k starts at 1 and goes up to n-1? Also with the function being 1/(2*(y-1))
How is (3/4)/[4((1-1)/4)^2] not 0 but 1/3 ??
In the example, where did the "after three years" info get applied?
It didn't, but it didn't need to. That was just setting up the question. It may as well have said; "after some period of time, let's call it 5 years..." or whatever. The actual mathematics question in there, if you want to call it that, is just that you have a random variable x, with the given MGF and you needed to calculate the expected value of g(x), where g(x) = 100(0.5)^x. That three years was just a random time, and was to set up the question and for it to have a tie to some real world situation. If you think about it, it wouldn't make sense for the value of the pice of equipment to be random at the present time... as they would know what it was worth then. Does that help?
Thank you, it really helps me.
Very clear! Thanks!
You forgot to link the proof. Otherwise, very clear.
By Taylor series expansion of e^(tX)!!!
Hi. What is the name of the whiteboard technology you use to solve problems?
You ARE the best! Thank you so much!
this vid was very helpful. thank you very much
very detailed explanation, thx a lot:)
Thank God for you
THANK YOU SIR
LIFE SAVER
nice staff... well understood
Why can't you differentiate the moment generating function simply and substitute t=0 to get the value of E(X)? THANKS
For the last question.
I notice the same
What would you then do with that E(X)? Notice we want the E(Y), which is the function we were given at the beginning, we can't just simply sub in that E(X)... Get it?
Thank you very much!
Really good video.. couple of queries, though.. a) what is the significance of moment, physical interpretation b) from where moment generating function originate? c) what if t assumes other numerical values? does it have any significance.. I would be glad if these are answered... TIA
Great questions! Moment generating functions are special cases of characteristic functions, which are Fourier transformations of probability density functions.
I may need to make another video just to answer your thoughtful questions.
Your questions also lead me to believe you are capable of getting some answers if I point you to a couple of resources:
From a historical perspective, the ideas of MGF were present in the book “Docrtine of Chances” by Abraham de Moivre (1667-1754). Mackey, George W. Harmonic analysis as the exploitation of symmetry-a historical survey. Bull. Amer. Math. Soc. 3 (1980), no. 1, part 1, 543-698 gives a great historical background of harmonic analysis and Fourier Transforms.
For an intuitive and visual description of Fourier transforms, I recommend a video by 3blue1brown: ruclips.net/video/spUNpyF58BY/видео.html. It explains that Fourier transforms decompose a signal into its components. It also does a good job of connecting to the idea of detecting the center of mass (c.f., the expected value can intuitively be thought as the center of mass of the probability density function).
A Statistical Path thank you very much sir, I will go through the resources...
Regards,
Abhisek
Thank you so much...it helped a lot.
The rewrite step you did in 14:56 was to obtain the exponencial, right? And you did that because you know you have the MGF
thank you so much for these videos
Thanks a lot, I'm subscribing.
where is the proof video? for E(x power k)
4:25
Thanks..from Thailand
very nice explanation! thanks!
You did not tell what 't' stands for here.
't' is just a dummy variable and has no meaning in life...acting as a label for the probability.
It doesn't matter. 't' can be anything really.
Good video! I have one question about your comment on calculating E(x).
Instead of using the MGF, is it possible to calculate E(x) directly. If possible, show me the detail please.
You just said that it is just difficult in the video.
Thank you in advance.
@16.47 why did we multiply the MGF by 100 rather than differentiate it?
Very nic e vid btw
Im very surprised to see that no one seems to provide a combinatorial derivation.
What's that?
Thank you.
Very good tutorial
1st in 2021💪
excelent video thanks alot
I wish you were my prob teacher.
This is amazing!
The mgf for a Uniform Discrete is not defined when t=0. I can still find moments by taking limit when t-->0 of the various derivatives. Is it actually incorrect when we say that the first moment is the first derivative evaluated at 0. Is it more correct to say that the first moment is the limit as t-->0 even if the mgf is not defined at t=0.
No.
And that distribution does not have an mgf, it doesn't exist.
where do i find the link to that proof again?
For the derivative of (3/4)/(1-(e^t/4)), I got a different answer of 3/16 instead of what the video shows. is anyone finding this?
www.wolframalpha.com/input/?i=0.75*%281-%28e%5Et%29%2F4%29%5E%28-1%29
but I haven't checked by hand
M'(x)=\frac{d}{dt}\left[ \frac{3}{4}(1-\frac{e^t}{4})^{-1}
ight]
=\frac{-3}{4}(1-\frac{e^t}{4})^{-2}\frac{-e^t}{4}
for t=0
M'(x)=\frac{-3}{4}(1-\frac{1}{4})^{-2}\frac{-1}{4}
=\frac{1}{3}
So wolfram was wrong and Actuarial Path is correct =D
tq
sir
It was quite helpful
As good as it gets!
I hope you explain me about Joint Moment Generating Functions , please
ZAAAAAAAAAMN
Hi first of all thanks a lot for your videos!
In the last example, why don't we take into consideration the 'after 3 year of use' ?
the function was in terms of three years.
@@fabianabrego6220 Thanks for this answer, was wondering the same then I re-read the question.
Thank you god (AKA "Actuarial Path")
Thank you. superb
am i the only one who realised the video was posted on valentines day? Is that why Statistic and I broke up? I wasnt as dedicated as this dude
Haha. I myself didn’t! Perhaps the reason I am still married lol
outstanding!
Anyone know where can I find derive joint mgf of bivariate normal distribution
What is 1-e^t/4 simplified to as one expression
Why does MGF start from the exponential function?
I am taking probability lecture at my school and professor gave an assignment but I couldn't solve 1 question from it. question askes that if X is a binomial random variable with parameter (n,p) and Y=X^2 . find the variance of Y by using characteristic function of X.
after watching this video I would solve this question in that way;
Var(Y) = E(Y^2) - [E(Y)]^2
Var(Y) = E[x^2)^2] - [E(x^2)]^2
after that point I thought that I need to take fourth derivative of the characteristic function of X and take second derivstive of characteristic function of X and take its square root am I right??
characteristic function of X = Φx(w) = (1 - p + pe^(jw))^n
I am not sure if this thought is correct can you help me?
I agree with your thought process, I think it's 100% correct.
احسنت موفق
thank you
quite helpful
.
it was awesome
thank you for the great video!
but the mic is just too much of a treble. With volume set to max at the beginning of video my ears pained and then I jumped from the seat.
Could you please fix and replace the audio if you have time. The crystal voice was a bit of a pain for my ears
Hello. Thank you for your amazing explanation.
I just have one question, when we do series expansion for Mx(t), how come it is E(tx + tx^1)?
If the function we're considering is e^(tx), how come it's not e^x and e^2x? Where does the exponential go?
Thank you :)
That is just the Maclaurin series of an exponential function.
where's the link of proving the relation between MGF and E[X]
Where's the proof?
Thanks🙏🏻
Love
THank you
Where is the proof pls , at 4:20
it is perfect until you say recall that a moment generating function is e(x),E(x^2)
Sir please explain how to solve Mx(ln.5)
thanks soo much !!!!!¬!!
Hello sir how can I get the link about deriving mgf
NICE