I’ve watched several videos from other content creators. Finally someone who actually knows the math and use cases and nuances rather than just someone going through some steps they learned about! This is what I have been searching for! Thank you and Thank you! I will now binge on your channel and recommend these to my nephew who is still in high school but is a math prodigy like his father was, who unfortunately passed away 10 years ago.
I usually don't comment but wow, this was just such a great video! Your explanation is both clear, simply explained and yet still in depth, this has really helped me!
This video was very well explained. I’m just confused about how to implement the replacement of the epsilon with the garch estimated residual. How can i isolate the epsilon? Is that just the actual value of the residual? I’m trying to implement in Python using pmdarima and archmodel. Thanks for your help.
thank you for your great video. a quick question. for a time series {X_t} follows ar(1)-garch(1,1) process, I agree that AR(1) process should be X_t = mu + a*X_t-1 + epsilon_t, but i don't understand why it is NOT X_t that follows the garch(1,1) process, i.e.: sigma_t^2 = omega + alpha * X_t-1^2 +beta * sigma_t-1^2. in other words, i don't know why it is the a_t, which equals sigma_t * epsilon_t(in your video), that follows garch process. I am really confuesd by this for a long time, thanks for your answer in advance.❤