🎉 🎉 RAPIDMINER 9.10 IS OUT!!! 🎉🎉
Download the latest version helping analytics teams accelerate time-to-value for streaming and IIOT use cases.
Optimal number for speculative rounds in forward selection
I am applying forward selection in developing SVM model. In the forward selection sub-process, I am applying 5 folds cross-validation. My questions are:
1. May I know what is the best / optimal number of speculative rounds in forward selection? Any articles that I can refer to as a guide?
It looks like the process non-stop when I use the default parameter, which the speculative round is set to 0, max no of attributes = 10, stopping behavior = without increase.
2. Based on the guide, if the speculative rounds is set to a value higher than 1, it will avoid getting stuck in the local optima. What does that means?
Hopefully the experts in Rapidminer could help me on this matter. Many thanks in advance.