基于密度聚类的网络数据流异常检测分析
Analysis of Network Data Flow Anomaly Detection Based on Density Clustering
高大拯1
作者信息
- 1. 中国科学技术大学计算机科学与技术学院,安徽 230026
- 折叠
摘要
阐述一种基于密度聚类的网络数据流异常检测算法,不仅能够适应数据样本分布的改变,还创新地提出一种双聚类图的算法,以避免在异常检测阶段,将潜在的微集群作为真正的异常值删除.这种技术可以显著提高检测性能,通过在多个数据集上的实验证明该方法的速度和有效性.
Abstract
This paper describes a density clustering based network data flow anomaly detection algorithm,which not only adapts to changes in the distribution of data samples,but also innovatively proposes a dual clustering class graph algorithm to avoid deleting potential micro clusters as true outliers during the anomaly detection stage.This technology can significantly improve detection performance,and its speed and effectiveness have been demonstrated through experiments on multiple datasets.
关键词
智能控制/网络数据流程/异常检测/密度聚类/流式数据Key words
intelligent control/network data flow/anomaly detection/density clustering/streaming data引用本文复制引用
出版年
2024
NETL
NSTL
万方数据