Hi, I’ve often come across this problem and have found genetic algorithms (GA) to be extremely useful. I wrote my first GA code in the 80’s and have extensive experience with the method. The package rgenoud is a very full featured GA implementation. Just code up your parameters as arguments to the function giving your method, random forests or whatever, then define a target variable for performance or fitness such as AUC or R^2, whatever is appropriate, and let the GA climb to the top of the fitness landscape. If you have a large problem you may want to speed things up by using parallel processes across cores or machines. Rgenoud handles that well.
Good luck! James > On Oct 11, 2019, at 4:21 PM, javed khan <javedbtk...@gmail.com> wrote: > > Hi > > I will appreciate if someone provide the link to some tutorials/videos > where parameters running are performed in R. For instance, if we have to > perform predictions/classification using random forest or other algorithm, > how different optimization algorithms tune the parameters of random forest > such as numbers of trees etc. > > Best regards > > [[alternative HTML version deleted]] > > ______________________________________________ > R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. > ______________________________________________ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
