Linear SVM

Is there any possibility to make an lineare SVM or a Soft Margin SVM instead of the Kernel SVMs you can make with the SVM Learner?

Hello @Student1988 -

Sorry for the delayed reply. The SVM Learner node, as you have noticed, only supports kernel methods. If you are able to translate your data into Spark (perhaps using KNIME’s extensions for Apache Spark and Local Big Data), you can use the Spark Linear SVM Learner node.

Of course you also have the option to use KNIME’s R Snippet and Python Snippet nodes to create such models using code.