基于深度学习的不均衡网络数据分类技术研究

在现实网络环境中,数据分布不均衡是普遍现象,也是研究的热点问题.利用传统机器学习算法解决该问题的研究成果较多,综述性研究也较丰富.但当前从深度学习的角度探讨数据不均衡问题已成为新趋势.对此,综述了基于深度学习方法的研究成果.通过对数据不均衡问题进行深入分析,从数据预处理、分类器设计及改进两大方面梳理相关技术路线,包括传...

Imbalanced Network Data Classification Technologies Based on Deep Learning

In the real network environment,imbalanced data distribution is a common phenomenon,and also a hot issue. Much research work has been done in using traditional machine learning algorithms to solve this problem. However,it has become a new trend to explore data imbalance from the perspective of deep learning. In this regard,the research work based on deep learning methods is summarized. Through in-depth analysis of data imbalance,the relevant technical routes are combed from two aspects of data preprocessing,classifier design and improvement,including the combination of traditional sampling methods and deep learning technologies,data synthesis using deep learning network models,cost sensitive learning,and end-to-end model design methods. Finally,open issues are proposed based on existing research.