Improved SMOTEBagging and its application in imbalanced data classification

Many real world data mining applications involve imbalanced data sets, When all kinds of data are unevenly distributed and the particular evens of interest may be very few when compared to the other class. Data sets that contain rare evens usually produces biased classifiers that have a higher predictive accuracy over the majority class, but poorer predictive accuracy over the minority class of interest. This paper presents a novel ensemble algorithm with improved SMOTE, and combines selective ensemble with Bagging, which balances the class distribution with improved SMOTEBagging algorithm. Experiments on four UCI data sets and protein-protein interaction experiments mentioned above prove the performance of the method.