ravshanbeksk

joined 11 months ago
[โ€“] ravshanbeksk@alien.top 1 points 11 months ago

Give an example of linear regression that does not support zero conditional mean assumption. Like auto regressive process. You cannot take derivative then because it is not differentiable. However gradient descent doesn't care about all that

[โ€“] ravshanbeksk@alien.top 1 points 11 months ago (1 children)

You can write a log likelihood function and take its derivative and set it equal to zero.