DE-LSSVM网络流量多分类方法
DE-LSSVM Method to Solve Multi-classification Problem of Network Traffic
徐轩 1姜志侠 1刘雪亚1
作者信息
- 1. 长春理工大学 理学院,长春 130022
- 折叠
摘要
支持向量机对网络流量进行分类时,支持向量机参数易导致分类模型的性能下降,分类精度低下等问题.针对该问题,提出一种改进差分优化算法与最小二乘支持向量机多分类器结合的方法,该模型采用具有自适应算子的DE算法作为优化方法,以LSSVM作为分类方法,交替进行,最终使分类结果最好.实验结果证明,该模型在网络流量多分类中,具有较低的均方根误差和更高的F1指数.
Abstract
When SVM classifies network traffic data,the support vector machine parameters tend to cause the perfor-mance of the classification model to decreas and the accuracy was low. For it,a method combining the improved differ-ential optimization algorithm with the least squares support vector machine was proposed. The model uses the DE algo-rithm with adaptive operator as the optimization method, and LSSVM was used as the classification method. These two methods were alternately running. The experimental results showed that the model had lower root mean square er-ror and higher F1 measure in the network traffic multi-classification.
关键词
网络流量分类/LSSVM分类模型/差分进化算法/多分类方法/参数选择Key words
network traffic classification/LSSVM classification model/differential evolution algorithm/multi-classifica-tion method/parameter selection引用本文复制引用
出版年
2019
NETL
NSTL
维普
万方数据