摘要
随着网络技术的快速发展,互联网中的安全问题愈发严重,因此定位互联网攻击者主机至关重要。目前的主机定位方法常常由于需要发送数据包和包含大量指纹信息而导致识别准确率过低。为了解决识别准确率过低缺点,文章首先介绍了常用的主机识别特征,其次分析主机识别中涉及的学习方法,最后针对主机识别过程中低识别率的问题提出了一种基于精简指纹的主机识别模型。该模型在支持向量机中使用卡林斯基-哈拉巴斯指数(Calinski-Harabaz Index,CHI)算法处理的特征进行预测,验证了该模型的精简指纹比完整指纹拥有更高的识别效率。
With the rapid development of network technology and the increasing security problems in the Internet,the task of host location of Internet attackers is crucial.Current host location methods often have low identification accuracy due to the need to send packets and the inclusion of large amounts of fingerprint information.To eliminate the impact of these drawbacks.This paper first introduces the commonly used host identification features,and then analyzes the learning methods involved in host identification.Finally,a host recognition model based on streamlined fingerprints is proposed to address the problem of low recognition rate in the host recognition process.The model predicts the features after passing the Calinski-Harabaz Index(CHI)algorithm using support vector machines,and verifies that the streamlined fingerprint of the model has better recognition efficiency than the complete fingerprint.
作者
魏建兵
WEI Jianbing(Gansu Forestry Vocational and Technical College,Tianshui Gansu 741020,China)
出处
《信息与电脑》
2023年第9期192-194,共3页
Information & Computer
基金
甘肃省重点人才项目“基于科研项目驱动化的创新型青年教师人才培育”(项目编号:2022RCXM001)。
关键词
网络安全
主机识别
特征选择
network security
host identification
feature selection