摘要
P2P网络入侵数据具有高维性、随机性以及复杂性等特征，会降低对其进行入侵检测的效率和稳定性。因此，本文提出一种基于PSO辨别树的P2P网络入侵检测方法，对P2P网络连接数据的主成分进行信息拟合，按照入侵数据特征的主成分构建PSO辨别树，获取同目标信息特征最为相似的估计关联特征；运用改进的BP神经网络对关联特征进行分析，最终可得到P2P网络入侵数据的特征信息，并通过无约束聚类关联方法优化获取的入侵数据特征，更好地完成对P2P网络信息安全的检测。实验结果表明，本文方法与其它网络入侵检测方法相比，具有较高的检测效率和准确率，取得了令人满意的效果，具有较大的发展前景和应用价值。
Abstract
Due to the P2P network intrusion data with high resistance, randomness and complexity characteristics, can re-duce the intrusion detection efficiency and stability. Therefore, this paper proPSOes a tree based on PSO discrimination of P2P network intrusion detection method, first to P2P network connection data of principal component information fit-ting, according to the data characteristics of the invasion of principal component build PSO discrimination tree, get with the target information features the most similar estimate correlation characteristic; Then using the improved BP neural network to correlation characteristics were analyzed, and finally can get P2P network intrusion data feature information, and through the unconstrained cluster correlation method for optimization of the invasion of the data characteristics, bet-ter complete the P2P network information security testing. The experimental results show that the method proPSOed in this paper and other network intrusion detection algorithm, has higher detection efficiency and accuracy, and achieved satisfactory effect, and it has great development prospects and application value.
NETL
NSTL
维普
万方数据