I haven't been able to sleep for days because I couldn't understand what I was learning and my exam's in 2 weeks. And within hours of just binge watching your videos and comparing them to my lecture notes, I am feeling super confident that I'm going to do well! Thank you, Ben!
@@sarikayadav6990 he just missed it plain and simple, when talking about population reg function, population parameters must be used - in that case, ui is error, when talking about sample reg fn, sample estimates must be used and ui should be treated as residuals
I don`t think residuals and error terms are treated the same. One is an observed one and the others are unobserved. When talking about the difference between the predicted data point and the actual data point, you should have used residual instead
Hi Ben! Great video. A quick question: If distribution of error term (u) can be any shape, including skewed one, how come we can maintain the assumption of E(u) = 0?
I haven't been able to sleep for days because I couldn't understand what I was learning and my exam's in 2 weeks. And within hours of just binge watching your videos and comparing them to my lecture notes, I am feeling super confident that I'm going to do well! Thank you, Ben!
This is a very well posed explanation of the concept.
Thank you
so clear! Thank you Ben, I just understand the part that i struggled for one semester! Cheers!!
This is a master piece
Now I understand what my professor meant when he said that we had to do some hypotesis about the distribution of the error terms. Pazzesco (Crazy)
Ben your literally an angel. This is extremely helpful!!
It was very helpful, thank you!
Is there a reason you use an s superscript when talking about the population, and not a p?
The s subscript is used for sample and not population. The data he considered here is of a sample. I hope it clears your doubt. :)
@@sarikayadav6990 he just missed it plain and simple, when talking about population reg function, population parameters must be used - in that case, ui is error, when talking about sample reg fn, sample estimates must be used and ui should be treated as residuals
how about if your trying to estimate the year when the population had reached a certain amount?
THE ONLY USEFUL VIDEO IN THE PLAYLIST for me
Yeah, for you.
I don`t think residuals and error terms are treated the same. One is an observed one and the others are unobserved. When talking about the difference between the predicted data point and the actual data point, you should have used residual instead
Hi Ben! Great video. A quick question: If distribution of error term (u) can be any shape, including skewed one, how come we can maintain the assumption of E(u) = 0?
A distribution can have mean zero and be skewed
Binge watching! My kind of Netflix
Great!
best!