# Relationship between structural or statistical properties and hardness of classification

I am trying to understand the relationship between structural or statistical properties of training dataset and hardness of classification in the context of binary classification with SVM using RBF kernel. I would like to predict the hardness without actually trying to classify the dataset. The obvious properties are the size of the dataset and features. What other meta-properties indicate that an SVM using RBF kernel will result in low accuracy and/or extremely expensive computation?

• I would look at the distribution of the distances. consider that the model depends not really on the features, but on the distance matrix. – carlo Sep 30 '20 at 16:47
• @carlo, could you clarify the definition of "distance"? May I know distance between what? – Omar Shehab Oct 1 '20 at 14:17
• RBF kernel uses distances $||x_1-x_2||$ – carlo Oct 1 '20 at 15:16

I know of one result on metalearning-like results for SVM: the mean distance between classes is a good heuristic value for the $$\gamma$$ hyperparameter for the RBF SVM (both behind paywall):