利用改进灰狼算法优化BP神经网络的入侵检测

神经网络技术被广泛应用于网络安全领域,在入侵检测中能够实现网络攻击的主动检测和攻击分类.然而随着恶意攻击的不断演化,神经网络技术存在的弊端日益显现.针对BP神经网络在入侵检测过程中存在的初始值随机性较大以及易陷入局部最优的问题,本文提出一种改进灰狼算法优化BP神经网络的入侵检测模型(IGWO-BP).首先,使用混沌映射初始化种群、设计非线性收敛因子以及动态权重策略对传统灰狼算法进行改进,并以此优化BP神经网络的初始权值和阈值,并运用改进BP神经网络对网络安全数据集进行实际检测.实验结果表明,IGWO-BP模型在NSL-KDD和UNSW-NB15数据集上取得了较优的检测结果,与其它现有模型相比性能也有较大提升.

Use Improved Grey Wolf Algorithm to Optimize BP Neural Network Intrusion Detection

Neural network technology is widely used in the field of network security.Intrusion detection can realize the active detection and classification of network attacks.However,with the continuous evolution of malicious attacks,the disadvantages of neural network technology have become increasingly apparent.Targeting BP neural networks in intrusion detection process,the initial value has large randomness and is easy to fall into a local optimal problem.This paper proposes an improved grey wolf algorithm to optimize the BP neural network intrusion detection model(IGWO-BP).First,the chaotic mapping is used to initialize the population,design nonlinear convergence factors and dynamic weighting strategies to improve the traditional grey wolf algorithm,and optimize the initial weights and thresholds of the BP neural network,and use the improved BP neural network to actually detect the network security data set.Experimental results show that the IGWO-BP model has achieved better detection results on the NSL-KDD and UNSW-NB15 datasets,and the performance has also been greatly improved compared to other existing models.