摘要
无线传感网络攻击流量类型较多,攻击流量检测方法难以满足多类型的流量数据,导致检测精度较差,为此提出基于TF-IDF算法的无线传感网络攻击流量检测方法。首先划分无线传感网络流量为连续型和离散型两类,采用独热编码处理连续型流量,归一化处理离散型流量;然后通过TF-IDF算法提取无线传感网络流量特征,利用特征向量集训练多通道自编码器,利用TF-IDF算法计算待检测的攻击流量数据特征在无线传感网络流量内出现的频率,以此对攻击流量进行排序;最后通过Softmax分类器输出最终流量类型检测结果。仿真结果表明,所提方法的检测精确度最低值为97.05%,虚警率最高值为2.01%、测试时间平均值为20.1 s,证明所提方法能高效、精确地实现无线传感网络攻击流量检测。
There are many types of attack traffic in wireless sensor networks,and the detection method of attack traffic is dfficult to meet multiple types of traffic data,resulting in poor detection accuracy.Therefore,a wireless sensor network attack traffic detection method based on TF-IDF algorithm is proposed.Firstly,the wireless sensor network traffic is divided into two type,i.e.,continuous and discrete.The continuous traffic is processed by using the unique heat coding,and the discrete traffic is normalized.Then,the traffic characteristics of the wireless sensor network are extracted by the TF-IDF algorithm,the multi-channel self-coder is trained by the feature vector set,and the TF-IDF algorithm is used to calculate the frequency of the attack traffic data characteristics to be detected in the wireless sensor net work traffic,so as to sort the attack traffic.Finally,the final traffic type detection result is output through Softmax classifier.The simulation results show that the minimum detection precision of the proposed method is 97.05%,the maximum false alarm rate is 2.01%,and the av erage test time is 20.1 s,proving that the proposed method can effectively and accurately detect attack traffic in wireless sensor networks.
作者
王晨
刘鑫
WANG Chen;LIU Xin(Department of Digital Information,Hebei Institute of International Business and Economics,Qinhuangdao Hebei 066311,China;School of Computer Science and Engineering,Central South University,Changsha Hu'nan 410000,China)
出处
《传感技术学报》
北大核心
2025年第4期744-748,共5页
Chinese Journal of Sensors and Actuators
基金
河北省教育厅项目(ZC2021217)。
作者简介
王晨(1982-),男,汉族,河北秦皇岛人,硕士,副教授。2004年于河北经贸大学获得学士学位,2010年于燕山大学获得硕士学位,现为河北对外经贸职业学院教师,主要从事计算机算法研究和计算机相关课程教学工作,wangchen202311@163.com;刘鑫(1982-),男,汉族,湖南衡阳人,博士研究生,副教授,高级工程师,研究方向:大数据、机器学习及自然语言处理等。
