基于密集连接卷积神经网络的入侵检测技术研究

Research on Intrusion Detection Technology Based on Densely Connected Convolutional Neural Networks

卷积神经网络在入侵检测技术领域中已得到广泛应用,一般地认为层次越深的网络结构其在特征提取、检测准确率等方面就越精确。但也伴随着梯度弥散、泛化能力不足且参数量大准确率不高等问题。针对上述问题,该文提出将密集连接卷积神经网络(DCCNet)应用到入侵检测技术中,并通过使用混合损失函数达到提升检测准确率的目的。用KDD 99数据集进行实验,将实验结果与常用的LeNet神经网络、VggNet神经网络结构相比。分析显示在检测的准确率上有一定的提高,而且缓解了在训练过程中梯度弥散问题。

Convolutional Neural Network(CNN)is widely used in the field of intrusion detection technology.It is generally believed that the deeper the network structure,the more accurate in feature extraction and detection accuracy.However,it is accompanied with the problems of gradient dispersion,insufficient generalization ability and low accuracy of parameters.In view of the above problems,the Densely Connected Convolutional Network(DCCNet)is applied into the intrusion detection technology,and achieve the purpose of improving the detection accuracy by using the hybrid loss function.Experiments are performed with the KDD 99 data set,and the experimental results are compared with the commonly used LeNet neural network and VggNet neural network structure.Finally,the analysis shows that the accuracy of detection is improved,and the problem of gradient vanishing during training is alleviated.