I did not understand why standard error is a good parameter! Let say the true mean is 85 and I have a sample of 83 and 83. Then unbiased estimation of mu is 83, unbiased estimation of variance is 0 which means that standard error of my estimation for mu, i.e 83 is 0.
If you check the normal distribution video in the playlist, you will see, the Gaussian PDF is not a closed form, so calculating the probability using integral over the PDF is not possible. There they have mentioned, how these statiscians used the standard normal distribution i.e N(mean = 0, variance = 1) and come with the CDF function which is the phi table. Now, computing the integral for any normal distribution is possible if we convert the given X into the standard normal value i.e Z = (x - mean)/std, and refer the phi table to get the probability. So, P(X
Very well presented and intuitive explanation of the CLT!
Please elaborate more on the examples.
20:20 anybody knows where this -1 came from ?
Phi(0) =0.5. So if you simplify then you would get -1
There is a property : Φ(x) + Φ(−x)=1. They replaced Φ(−x) with 1 - Φ(x).
VERY GOOD 👍
This lecture had a lot of gaps.
I did not understand why standard error is a good parameter! Let say the true mean is 85 and I have a sample of 83 and 83. Then unbiased estimation of mu is 83, unbiased estimation of variance is 0 which means that standard error of my estimation for mu, i.e 83 is 0.
OH, I got the answer in the end of the lecture at the time 47:00. This estimation is valid under i.i.d assumption.
What is this phi that is being used here?
Capital phi is the CDF of standard normal
If you check the normal distribution video in the playlist, you will see, the Gaussian PDF is not a closed form, so calculating the probability using integral over the PDF is not possible. There they have mentioned, how these statiscians used the standard normal distribution i.e N(mean = 0, variance = 1) and come with the CDF function which is the phi table. Now, computing the integral for any normal distribution is possible if we convert the given X into the standard normal value i.e Z = (x - mean)/std, and refer the phi table to get the probability.
So, P(X