what's the best alternative in R for variable selection? I have more than 300 predictor variables for classification issue using machine learninig algorithms (random forest , SVM, CART).
Benjamin, could you give the reference you stated above about SVM and CART variable selection?
Also, you can read the article of Guyon & Elisseeff: An Introduction to Variable and Feature Selection (http://www.jmlr.org/papers/volume3/guyon03a/guyon03a.pdf). You can then search for a good package of the method you prefer. Meta-heuristic search algorithms are other options.
You can see my article with a classification application and several R packages' information: Article Improving the performance of statistical learning methods wi...
Since you have around 300 variables to analyze using R, I will advise you use WEKA to select the best features prior to analysis using R as WEKA has implemented functions for features selection like Best First Search, Greedy search, Genetic search and so on.
My favorites are lasso/elastic net and gradient boosting. See the attached for why and full details. Best wishes, David
https://www.amazon.com/Everyone-Advanced-Analytics-Graphics-Addison-Wesley/dp/013454692X/ref=sr_1_1?ie=UTF8&qid=1522952612&sr=8-1&keywords=lander+jared
- Badges
- Science topic
- Similar topics
- Clinical Pharmacology
- Dosing