基于深度学习的入侵检测模型

Intrusion Detection Model Based on Deep Learning

针对网络流量数据具有空间和时间的双重特征,提出了一种基于深度学习的入侵检测模型。首先,通过二分支卷积神经网络提取网络流量数据的空间特征,利用其分支结构的特点使得不同的卷积层对同一个数据样本进行粗化提取和细化提取,既保留了数据的总体特征,又从低级特征中迭代提取出更复杂的特征;然后,利用门控循环单元网络顺序敏感性的优势,挖掘网络流量数据的时序特征;最后,使用KDDCUP99数据集对入侵检测模型进行训练、验证和测试。实验结果表明,与传统的基于机器学习的模型相比,该模型具有更高的检测准确率。

Aiming at the dual characteristics of space and time in network traffic data, an intrusion detection model based on deep learning is proposed. Firstly, the spatial characteristics of network traffic data are extracted by two branch convolution neural network, and the characteristics of its branch structure are used to make different convolution layers roughen extract and refine extract the same data sample, which not only preserves the overall characteristics of the data, but also extracts more complex characteristics from the low-level characteristics iteratively. Then, the time characteristics of network traffic data are mined by taking the advantage of sequential sensitivity of gated loop unit network. Finally, KDDCUP99 data set is used to train, verify and test the intrusion detection model. The experimental results show that compared with the traditional model based on machine learning, this model has higher detection accuracy.