关键词: K近邻, 径向基过采样, 样本势, 自然近邻, 差分进化, 不平衡数据分类

Abstract: In dealing with the problem of imbalanced data classification, oversampling methods are considered effective strategies. Existing methods mostly employ K-nearest neighbor (KNN) technique to select oversampling seed samples, but changes in KNN parameter values often lead to significant instability for most oversampling methods. The recently proposed radial-based oversampling (RBO) method can address this issue, but it tends to introduce a substantial amount of noise after oversampling. In this paper, we propose an imbalanced data oversampling method based on sample potential and noise evolution to further iteratively refine the oversampled dataset. The core steps are as follows: Firstly, the RBO method is used to synthesize minority class samples and improve the imbalance of the original data by calculating sample potential. Secondly, natural neighbors (NaN) is employed as an error detection technique to identify suspected noise samples in the oversampled dataset. Finally, an improved differential evolution (DE) method is applied to iteratively refine the detected suspected noise samples. Compared to traditional oversampling methods, the proposed method can better explore important boundary information in the dataset, thus providing more assistance to classifiers to improve their classification performance. Extensive comparative experiments were conducted on 22 benchmark datasets with seven classical sampling methods (combined with three different classifiers). The experiment results show that the proposed method achieves higher F1 and G-mean values and is superior in noise handling compared to sampling methods with post-filters, which can more effectively deal with the problem of imbalanced data classification. In addition, statistical analysis also indicates a higher Friedman Ranking for the proposed method.

Key words: K-nearest neighbor, radial-based oversampling, sample potential, natural neighbor, differential evolution, imbalanced data classification