摘要
为了解决物联网终端接入设备较多时,接入结果可靠性较低的问题,提出基于改进机器学习算法的物联网终端安全接入方法。依据物联网终端设备的标识性和可采集性原则,使用SDK接口提取物联网设备信息;依据该信息对相对应的物联网终端设备流量进行编码和特征提取,生成设备指纹信息;构建物联网终端设备指纹信息库,将物联网终端设备指纹作为输入,使用改进朴素贝叶斯算法优化匹配设备指纹信息库,若匹配成功则允许其接入,反之则拒绝其接入。实验结果表明,该方法P-R曲线的平衡点对应数值分别在0.88和0.83左右,接入结果可靠性较高,可有效实现物联网终端设备负样本匹配,保障设备的安全接入。
In order to solve the problem of low reliability of access results when there are many terminal access devices,a secure access method of IoT terminals based on improved machine learning algorithm is proposed.According to the principle of identification and collectability of IoT terminal equipment,after extracting the information of IoT terminal equipment by using SDK interface,we encode and extract the corresponding traffic of IoT terminal equipment according to the information,generate equipment fingerprint information and build the fingerprint information database of IoT terminal equipment.The fingerprint of IoT terminal equipment is taken as the input,the improved naive Bayesian algorithm is used to optimize the fingerprint information base of matching equipment.If the matching is successful,it is allowed to access,otherwise it is rejected to realize the secure access process of IoT terminal.The experimental results show that the corresponding values of the equilibrium point of the P-R curve of this method are about 0.88 and 0.83 respectively,and the reliability of the access results is high,which can effectively realize the negative sample matching of IoT terminal equipment and ensure the safe access of equipment.
作者
景钰文
韩世海
朱珠
JING Yuwen;HAN Shihai;ZHU Zhu(Electric Power Science Research Institute of State Grid Chongqing Electric Power Company,Chongqing 401123,China)
出处
《微型电脑应用》
2024年第1期141-144,共4页
Microcomputer Applications
关键词
改进机器学习
物联网
终端安全接入
朴素贝叶斯
设备指纹
improving machine learning
Internet of Things
secure terminal access
naive Bayes
device fingerprint
