I am applying SVM with RBF kernel on binary imbalanced problems.
I would like to achieve better performance by optimizing penalty and sigma parameters. I already saw this post: https://tech.knime.org/forum/knime-general/svm-lerner-optimization and some other but they were not really helpful.
I downloaded extension with Parameter Optimization Loop Start/End.
Could you please explain me how to optimize SVM parameters in Cross Validation meta nodes, KNIME 3.1.2?
Cross Validation is nto the same as parameter optimization. For parameter optimization you can take a look at this workflow: https://www.knime.org/nodeguide/analytics/optimization/parameter-optimization (You can download it from our example server)
It will iterate through all possible combinations of your parameters.
Does this help? I also attached you a small example workflow to get started.
You might also want to see our looping videos, which are very helpful, to get the idea of loops. https://tech.knime.org/node/58953/view