Yet another problem with generating synthetic example between the minority seed and its majority neighbour is to direct its position rather closer to the minority class than to the majority class represented by n.Similarly to solutions from Borderline-SMOTE2 methods we want to restrict the range of interval where the new example could be randomly located.So, in some cases of safe levels we do not consider right boundary of the interval as 1 but as threshold < 1.
正在翻譯中..
