News for hackers. Stuff that matters.
Re:An Introduction to Gradient Descent and Linear Regression by mattnedrich on Thursday June 26, 2014 @02:34PM (#7950945)
Author here. These are all great points. I chose linear regression because it's simple and easy to understand. There is obviously a simple closed form solution to this problem. However, as already mentioned, gradient descent has some nice advantages like scaling. Further, a "close enough" solution/approximation might be sufficient depending on the application.
Re:An Introduction to Gradient Descent and Linear Regression by grayclhn on Thursday June 26, 2014 @03:20PM (#7951161)
There are algorithms to add variables and observations to an estimated linear regression model, so scaling is not an advantage of gradient descent. You should mention in the post that there are better ways to estimate this model, even if you're just presenting it as an example.
Re:An Introduction to Gradient Descent and Linear Regression by nkurz on Thursday June 26, 2014 @03:28PM (#7951195)
One tiny improvement to the article would be to change the "visualization of the surface" images to to show the part of the surface where the solution is eventually found. The 3D graph cuts off at y = -2, which makes things a little harder to understand.