A Novel Neighborhood-Weighted Sampling Method for Imbalanced Datasets

扫码查看

原文链接

万方数据

外文摘要：The weighted sampling methods based on k-nearest neighbors have been demonstrated to be ef-fective in solving the class imbalance problem.However,they usually ignore the positional relationship between a sample and the heterogeneous samples in its neighbor-hood when calculating sample weight.This paper pro-poses a novel neighborhood-weighted based sampling method named NWBBagging to improve the Bagging al-gorithm's performance on imbalanced datasets.It con-siders the positional relationship between the center sample and the heterogeneous samples in its neighbor-hood when identifying critical samples.And a parameter reduction method is proposed and combined into the en-semble learning framework,which reduces the paramet-ers and increases the classifier's diversity.We compare NWBBagging with some state-of-the-art ensemble learn-ing algorithms on 34 imbalanced datasets,and the result shows that NWBBagging achieves better performance.

外文关键词：

Ensemble learningClass imbalanceWeighted samplingData mining

作者：

GUANG Mingjian、YAN Chungang、LIU Guanjun、WANG Junli、JIANG Changjun

展开 >

作者单位：

Key Laboratory of Embedded System and Service Computing,Tongji University,Ministry of Education,Shanghai 201804,China

National(Province-Ministry Joint)Collaborative Innovation Center for Financial Network Security,Tongji University,Shanghai 201804,China

基金：

国家重点研发计划

项目编号：

2018YFB2100801

出版年：

2022

DOI：

10.1049/cje.2021.00.121

电子学报(英文)

CSTPCDSCIEI

ISSN：1022-4653

年,卷(期)：2022.31(5)

参考文献量1