您的位置:首页  > 论文页面

一种基于C4.5决策树的网络流量分类特征选择方法

发表时间:2015-03-31  浏览量:1558  下载量:701
全部作者: 姜羽,华俊,胡静,宋铁成,刘世栋,郭经红
作者单位: 东南大学移动通信国家重点实验室;中国电力科学研究院信息通信研究所
摘 要: 分析特征选择及其在网络流量分类技术中的应用,提出一种基于C4.5决策树算法的特征选择方法,并利用Moore数据集进行实验验证。结果表明,所提方法可以在不影响分类准确率的前提下有效降低特征维数,减少建模和分类时间。
关 键 词: 通信网络技术;特征选择;C4.5算法;建模时间
Title: A feature selection method of network traffic classification based on C4.5 decision tree
Author: JIANG Yu, HUA Jun, HU Jing, SONG Tiecheng, LIU Shidong, GUO Jinghong
Organization: National Key Laboratory of Mobile Communications, Southeast University; Research Institute of Information Technology & Communication, China Electric Power Research Institute
Abstract: This paper analysed the feature selection and its application in network traffic classification technology, and then proposed an application of feature selection which fitted C4.5 decision tree algorithm verified using the Moore dataset. The results show thatthe proposed method can reduce the dimension of features and decrease modeling and classification time without affecting the classification accuracy.
Key words: communication network technology; feature selection; C4.5 algorithm; modeling time
发表期数: 2015年3月第6期
引用格式: 姜羽,华俊,胡静,等. 一种基于C4.5决策树的网络流量分类特征选择方法[J]. 中国科技论文在线精品论文,2015,8(6):564-572.
 
0 评论数 0
暂无评论
友情链接