SVM C Parameter. higher C values allow for 'softer' boundaries
I am taking rapid miner analyst certification in a couple of days.
In Rapid Miner documentation it says, higher C values allow for 'softer' boundaries and lower values create 'harder' boundaries. Is this correct ? Ideally large C values should lead to hard margins right ? Pls Suggest
Answers
-
Hi,
I guess this depends a bit on how you define the terms "hard" and "soft" here :-)
Think about a RBF kernel which can be non-linear. "Soft" in the documentation means that you can "bend" or "move" the hyperplace more, i.e. you can bring it into a shape which reduces the training error. Which is exactly what happens for higher values of C. While a "hard" boundary tends to stay in its shape and cannot be moved that easily. This can therefore increase the training error.
I would probably somewhat avoid to use those fuzzy terms altogether now but I am pretty sure that is what is meant by the documentation.
Hope this helps,
Ingo
2
