A hybrid evolutionary preprocessing method for imbalanced datasets

Abstract

Imbalanced datasets are commonly encountered in real-world classification problems. Many machine learning algorithms are originally designed for well-balanced datasets, therefore re-sampling has become an important step to pre-process imbalanced data. This aims to balance the datasets by increasing the samples of the smaller class or decreasing the samples of the larger class, which are known as over-sampling and under-sampling, respectively. In this paper, a sampling strategy that is based on both over-sampling and under-sampling is proposed, in which the new samples of the smaller class are created based on fuzzy logic. Improvement of the datasets is done by the evolutionary computational method of Cross-generational elitist selection, Heterogeneous recombination and Cataclysmic mutation (CHC) that under-samples both the minority and majority samples. Consequently, a hybrid preprocessing method is proposed to re-sample imbalanced datasets. The evaluation is done by applying the Support Vector Machine (SVM), C4.5 decision tree and nearest neighbor rule to train a classification model from the re-sampled training sets. From the experimental results, it can be seen that our proposed method im- proves both the F −measure and AUC. The over-sampling rate and complexity of the classification model are also compared. Our proposed method is found to be superior to all other methods under comparison and it is more robust in different classifiers.