Let's say you were given a task : Make me the MLE (Maximum Likelihood Est) of: Y=f(xi) + ei (as presented in the vid) and they do not specify any kernel or how the f(xi) looks like. How to attack this?
At the beginning of the video, you define that the variance is constant for all e_i. Worse in the example you give, it is clear that the variance is not constant
Is it possible to get a playlist to reflect the sequence of the videos in the series. Makes it easier to download the entire series in one step.
Thank you so much Brian, this is super helpful!!!!
love your video! love your voice!
Thank you for this great video. The references are very helpful.
this is for me but I am understanding you . thank you
Highly appreciated!
so much underrated! Thanks Brian :)
Thanks Brian for informative video. Is there a way to apply differentiation post smoothing..?
I think Epenechnikov kernel should be in form of (a+bx^2). you accidentally put the square outside the brackets
Let's say you were given a task : Make me the MLE (Maximum Likelihood Est) of:
Y=f(xi) + ei (as presented in the vid)
and they do not specify any kernel or how the f(xi) looks like.
How to attack this?
is not the Epanechnikov kernel defined as 3/4 (1 - x^2) not 3/4 (1-x)^2
Great job. Can you Share the slides?
Is kernel regression is another name of kernel ridge regression?
They aren’t equivalent. “Ridge” regression provides a different way of balancing the bias-variance tradeoff.
At the beginning of the video, you define that the variance is constant for all e_i. Worse in the example you give, it is clear that the variance is not constant
ty