Lasso function behaving completely differently on 2014b and 2017b
Show older comments
I'm using the lasso function from the statistics and machine learning toolbox for a feature selection algorithm on 2014b and 2017b and getting different results. I can't really share details/code because I'm using this for part of a research project. The 2014b version produces a much sparser set of of coefficients than the 2017b version (same exact code). This results in better performance from the 2014b version. One difference is the 2014 version uses just the statistics toolbox. Are there any notable changes in the algorithm between versions? Are there any parameters I need to use to get similar behavior?
Answers (0)
Categories
Find more on Linear Regression in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!