网络窃密木马的自适应检测模型研究

以窃密木马的网络通信行为为研究对象, 旨在建立具备自适应动态反馈学习能力的窃密木马检测系统。为实现具备自适应反馈能力的窃密木马检测系统, 研究了网络数据流及窃密木马通信数据流的特征, 引入基于增量学习的支持向量机分类算法, 提出了增量学习样本的构建方法, 建立动态反馈木马检测模型, 设计并实现了具备自适应能力的窃密木马检测系统。真实网络环境下的实验结果验证了系统的有效性和高效性。当网络环境发生变化时, 系统体现了良好的适应性。

Research on adaptive network theft Trojan detection model

This paper intended to establish theft Trojans detection system capable of adaptive dynamic feedback learning. To achieve this goal, it first studied the characteristics of the network data stream and theft Trojans communication data stream. Then it introduced support vector machine algorithm based on incremental learning, and proposed the construction method of incremental learning samples, designed and implemented the theft Trojan detection system with self-adaptive capability. Finally, the experimental results in a real network environment verify the effectiveness and efficiency of the system.