ALBERT

Description: Background: Over-sampling methods based on Synthetic Minority Over-sampling Technique (SMOTE) have beenproposed for classification problems of imbalanced biomedical data. However, the existing oversamplingmethods achieve slightly better or sometimes worse result than the simplest SMOTE. In orderto improve the effectiveness of SMOTE, this paper presents a novel over-sampling method usingcodebooks obtained by the learning vector quantization. In general, even when an existing SMOTEapplied to a biomedical dataset, its empty feature space is still so huge that most classification algorithmswould not perform well on estimating borderlines between classes. To tackle this problem, ourover-sampling method generates synthetic samples which occupy more feature space than the otherSMOTE algorithms. Briefly saying, our over-sampling method enables to generate useful syntheticsamples by referring to actual samples taken from real-world datasets. Results: Experiments on eight real-world imbalanced datasets demonstrate that our proposed over-samplingmethod performs better than the simplest SMOTE on four of five standard classification algorithms.Moreover, it is seen that the performance of our method increases if the latest SMOTE called MWMOTEis used in our algorithm. Experiments on datasets for ß-turn types prediction show someimportant patterns that have not been seen in previous analyses. Conclusions: The proposed over-sampling method generates useful synthetic samples for the classification of imbalancedbiomedical data. Besides, the proposed over-sampling method is basically compatible withbasic classification algorithms and the existing over-sampling methods.