摘要
传统机器学习从原始网络游戏流量准确提取特征面临巨大的困难。目前,已有研究者使用深度学习从原始流量中自动化提取特征,然后再进行流量分类。论文针对深度学习模型开销较大,计算复杂度较高等问题,提出了一种基于PCA和多分类SVM的网络游戏流量识别模型。首先将游戏流量预处理为28×28的灰度图,然后利用主成分分析PCA进行特征降维,最后基于多分类SVM算法进行游戏流量的分类和识别。经过三折交叉实验,结果表明在保持较好的分类效果基础上能够实现较低的计算复杂度。
Traditional machine learning is facing great difficulties in accurately extracting features from raw online game traffic.At present,some researchers have used deep learning to automatically extract features from raw traffic and then classify traffic.This paper aims at the problems of high overhead and high computational complexity of deep learning model,an online game traffic recognition model based on PCA and multi-classification SVM is proposed.Firstly,the game traffic is preprocessed into a 28×28gray scale map.Then,principal component analysis(PCA)is used for feature dimension reduction.Finally,the game traffic is classified and recognized based on multi-classification SVM algorithm.The results of three-fold crossover experiment show that the algorithm can achieve lower computational complexity while maintaining better classification effect.
作者
宁安安
年梅
张俊
NING Anan;NIAN Mei;ZHANG Jun(School of Computer Science and Technology,Xinjiang Normal University,Urumqi 830054;Xinjiang Institute of Physics and Chemistry Technology,Chinese Academy of Sciences,Urumqi 830011)
出处
《计算机与数字工程》
2024年第9期2739-2744,共6页
Computer & Digital Engineering
基金
自治区高校科研项目(编号:XJEDU2017S032)
国家重点研发计划子课题(编号:E1182101)资助。
关键词
机器学习
PCA
多分类SVM
流量识别
深度学习
machine learning
PCA
multi-classification SVM
traffic identification
deep learning
