Related to your question:
1 Like
If anyone is interested, i have implemented a faster and also numerically stable method for underdetermined linear systems in this post:
2 Likes
Just to mention that I wrote a simple package to do what I was searching for when I found this thread, which was a very simple interface to do simple fits of 2D data:
8 Likes
I also needed basic linear regression, a bit beyond X \ y, and didn’t want the full GLM dependency chain, so I tried to incorporate the various suggestions from this thread in a very simple, minimal-dependency package, see https://github.com/st--/LinearRegression.jl - in case anyone else who stumbles across this thread may find it useful!
I also put links to various packages for more fancy versions of linear regression (generalized, ridge, sparse, bayesian, online, …) into the readme.
7 Likes
great stuff, I just walked the same path
1 Like