面向不平衡数据分类的高维超球体过采样方法

林静怀; 刘治宇; 李军良; 高欣; 李泽科; 唐志军; 余斯航; 徐建航

Chinese

您当前的位置：

首页 >

文章列表页 >

面向不平衡数据分类的高维超球体过采样方法

更新时间：2026-09-04

- 面向不平衡数据分类的高维超球体过采样方法
- High-dimensional hypersphere oversampling method for imbalance classification
- 微电子学与计算机 2021年38卷第5期页码：65-72
- 作者机构：
  
  1. 国网福建省电力有限公司,福建,福州,350003
  2. 北京邮电大学人工智能学院,北京,100876
  3. 南瑞集团(国网电力科学研究院机构 )有限公司,北京,100192
  4. 国网福建省电力有限公司电力科学研究院,福建,福州,350007
- 作者简介：
- 基金信息：
  
  国网福建省电力有限公司总部科技项目 52130419002F
- DOI：
  中图分类号：
- 纸质出版：2021
- 稿件说明：
移动端阅览
林静怀, 刘治宇, 李军良, 高欣, 李泽科, 唐志军, 余斯航, 徐建航. 面向不平衡数据分类的高维超球体过采样方法[J]. 微电子学与计算机, 2021, 38(5): 65-72.

LIN Jing-huai, LIU Zhi-yu, LI Jun-liang, et al. High-dimensional hypersphere oversampling method for imbalance classification[J]. 2021, 38(5): 65-72.
林静怀, 刘治宇, 李军良, 高欣, 李泽科, 唐志军, 余斯航, 徐建航. 面向不平衡数据分类的高维超球体过采样方法[J]. 微电子学与计算机, 2021, 38(5): 65-72. DOI：

LIN Jing-huai, LIU Zhi-yu, LI Jun-liang, et al. High-dimensional hypersphere oversampling method for imbalance classification[J]. 2021, 38(5): 65-72. DOI：

摘要

在机器学习不平衡分类方法研究中，由于多数类与少数类样本数量之间存在较大差异，导致分类器易出现判定准确率低的问题.以SMOTE为代表的一类过采样方法是处理该问题的一种有效手段.该类方法在选定的线段中随机生成少数类新点来重新平衡数据集，但存在忽略少数类样本在超维空间中分布多样性的缺陷.本文提出一种面向不平衡数据分类的高维超球体过采样(HS-SMOTE)方法.在少数类样本集上通过随机抽样获得需要平衡的样本数目，在此基础上依次对每一样本通过欧氏距离选取其在少数类分布空间中的对应最近邻点，以两点连线中点为球心在超维空间构建采样超球体，在此区域内通过维度空间距离迭代随机生成所需的少数类新点，在类别样本数据再平衡的基础上增加少数类样本的空间分布多样性.在15组KEEL不平衡数据集上结合随机森林(RF)分类器开展了大量实验，与6种典型过采样方法相比，所提方法在G-mean以及F1-score指标上均有较好的表现，并通过了2种统计学假设检验方法的有效性验证.

Abstract

In the research of imbalance classification methods of machine learning

the classifier is prone to the problem of low judgment accuracydue to the large difference of the number between the majority class and the minority class. A class of oversampling methods represented by SMOTE are effective to deal with this problem. These types of methods randomly generate the minority new points in the selected line segment to rebalance the data set

but there is the defect of ignoring the diversity of minority samples in the super-dimensional space. A high-dimensional Hypersphere-SMOTE (HS-SMOTE) method isproposed for imbalanced data classification. On the minority sample set

the number of samples that need to be balanced is obtained by random sampling

and based on this sampling

for each sample

its corresponding nearest neighbor is selected in turn through the Euclidean distance in the minority distribution space

and the midpoint of the two points is used for the center to construct a sampled hypersphere in the super-dimensional space. In this area

randomly generate the required minority new points through the dimensional space distance iteration

thus the spatial distribution diversity of the minority samples is increased on the basis of rebalancing the category sample data. A large number of experiments have been carried out on 15 sets of KEEL imbalanced data sets combining Random Forest (RF)classifiers. Compared with the 6 typical oversampling methods

the method proposed in the article has good performance on G-meanandF1-score indicators

and have passed the validity verification of two statistical hypothesis testing methods.

关键词

Keywords

references

GALAR M, FERNANDEZ A, BARRENECHEA E, et al. A review on ensembles for the class imbalance problem: bagging-, boosting-, and hybrid-based approaches[J]. IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews), 2012, 42(4): 463-484. DOI:10.1109/TSMCC.2011.2161285.

HAN H, WANG W Y, MAO B H. Borderline-SMOTE: a new over-sampling method in imbalanced data sets learning[C]//Proceedings of 2005 International Conference on Intelligent Computing. Hefei, China: Springer, 2005: 878-887.

DOUZAS G, BACAO F. Self-Organizing Map Oversampling (SOMO) for imbalanced data set learning[J]. Expert Systems with Applications, 2017(82): 40-52. DOI:10.1016/j.eswa.2017.03.073.

HE H B, BAI Y, GARCIA E A, et al. ADASYN: adaptive synthetic sampling approach for imbalanced learning[C]//Proceedings of 2008 IEEE International Joint Conference on Neural Networks(IEEE World Congress on Computational Intelligence). Hong Kong, China: IEEE, 2008. DOI: 10.1109/IJCNN.2008.4633969.

DOUZAS G, BACAO F, LAST F. Improving imbalanced learning through a heuristic oversampling method based on k-means and SMOTE[J]. InformationSciences, 20 18(465): 1-20. DOI: 10. 10 16/j.ins.20 18.06.056.

DEWAN I, RAO B L S P. Wilcoxon-signed rank test for associated sequences[J]. Statistics % Probability Letters, 2005, 71(2): 131-142. DOI:10.1016/j.spl.2004.10.034.

贾统, 李影, 吴中海. 基于日志数据的分布式软件系统故障诊断综述[J]. 软件学报, 2020, 31(7): 1997-2018. DOI:10.13328/j.cnki.jos.006045. JIA T, LI Y, WU Z H. Survey of state-of-the-art log-based failure diagnosis[J]. Journal of Software, 2020, 31(7): 1997-2018. DOI:10.13328/j.cnki.jos.006045.

BREIMAN L. Random forests[J]. Machine Learning, 2001, 4 5(1): 5-32. DOI:10.1023/A:1010933404324.

ELREEDY D, ATIYA A F. A comprehensive analysis of Synthetic Minority Oversampling Technique (SMOTE) for handling class imbalance[J]. Information Sciences, 2019, 505: 32-64. DOI:10.1016/j.ins.2019.07.070.

李睿峰, 许爱强, 孙伟超, 等. 基于样本重采样的电路非平衡数据预处理方法[J]. 系统工程与电子技术, 2020, 42(11): 2654-2660. DOI:10.3969/j.issn.1001-506X.2020.11.30. LI R F, XU A Q, SUN W C, et al. Preprocessing method based on sample resampling for imbalanced data of electronic circuits[J]. Systems Engineering and Electronics, 2020, 42(11): 2654-2660. DOI:10.3969/j.issn.1001-506X.2020.11.30.

GARCíA S, FERNÁNDEZ A, LUENGO J, et al. Advanced nonparametric tests for multiple comparisons in the design of experiments in computational intelligence and data mining: experimental analysis of power[J]. Information Sciences, 2010, 180(10): 2044-2064. DOI:10.1016/j.ins.2009.12.010.

LI J Y, FONG S, WONG R K, et al. Adaptive multi-objective swarm fusion for imbalanced data classification[J]. Information Fusion, 20 18(39): 1-24. DOI: 10. 10 16/j.inffus.20 17.03.007.

CHAWLA N V, BOWYER K W, HALL L O, et al. SMOTE: synthetic minority over-sampling technique[J]. Journal of Artificial Intelligence Research, 2002, 16(1): 321-357. DOI:10.1613/jair.953.

LIANG X W, JIANG A P, LI T, et al. LR-SMOTE -An improved unbalanced data set oversampling based on K-means and SVM[J]. Knowledge-Based Systems, 2020, 196: 105845. DOI: 10.1016/j.knosys.2020. 105845.

李新鹏, 高欣, 何杨, 等. 不平衡数据集下基于自适应加权Bagging-GBDT算法的磁盘故障预测模型[J]. 微电子学与计算机, 2020, 37(3): 14-19. DOI:10.19304/j.cnki.issn1000-7180.2020.03.004. LI X P, GAO X, HE Y, et al. Prediction model of disk failure based on adaptive weighted bagging-GBDT algorithm under imbalanced dataset[J]. Microelectronics % Computer, 2020, 37(3): 14-19. DOI:10.19304/j.cnki.issn1000-7180.2020.03.004.

RAGHUWANSHI B S, SHUKLA S. Class imbalance learning using UnderBagging based kernelized extreme learning machine[J]. Neurocomputing, 2019(329): 172-187. DOI: 172

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于计算机视觉的电子元器件表面缺陷检测综述

混沌和柯西变异的蝗虫优化算法及特征选择

融合特征选择的Android恶意逃避攻击研究

一种多X值输入下测试覆盖率损失的预测方法

基于量子计算的分类和聚类算法综述