I would like to adopt a general strategy for dealing with an very unbalanced dataset, where my "positive" group corresponds to 1/40 of all the observations. The reason why I ask it is because all the machine learning methods will decrease overall classification error by misclassifying the positive group, which is what is really of interest to me.
Is there any general technique for dealing with this? The technique must be useful across many machine learning methods for classification, like Random Forest, adaboosting, SVM, etc.
