面向入侵检测系统的Deep Belief Nets模型

Deep belief nets model oriented to intrusion detection system

连续的网络流量会导致海量数据问题,这为入侵检测提出了新的挑战。为此,提出一种面向入侵检测系统的深度信念网络(deep belief nets oriented to the intrusion detection system,DBN-IDS)模型。首先,通过无监督的、贪婪的算法自底向上逐层训练每一个受限玻尔兹曼机(restricted Boltzmann machine,RBM)网络,使得大量高维、非线性的无标签数据映射为最优的低维表示;然后利用带标签数据被附加到顶层,通过反向传播(back propagation,BP)算法自顶向下有监督地对RBM网络输出的低维表示进行分类,并同时对RBM网络进行微调;最后,利用NSLKDD数据集对模型参数和性能进行了深入的分析。实验结果表明,DBN-IDS分类效果优于支持向量机(support vector machine,SVM)和神经网络(neural network,NN),适用于高维、非线性的海量入侵数据的分类处理。

It puts forward a new challenge with intrusion detection, the continuous collection of traffic databy the network leads to the massive data problems. Therefore, a deep belief nets model oriented to the intrusiondetection system （DBN-IDS） is proposed. First, an unsupervised, greedy algorithm is employed to train eachrestricted Boltzmann machine （RBM） at a time by a bottom-up approach, which makes large amounts of nonlinearhigh-dimensional unlabeled input data can be sampled as optimal low-dimensional feature representations.Second, using the labeled data at the top layer, the supervised back propagation （BP） algorithm is employed inclassifying the learned low-dimensional representations and fine-tuning the RBM network simultaneously. Theparameters and the performance of the model are deeply analyzed on NSL-KDD dataset. Experimental resultsdemonstrate that the DBN-IDS outperforms the support vector machine （SVM） and neural network （NN） , andwhich is a feasible approach in intrusion classification for the high-dimensional, nonlinear and large-scale data.