论文部分内容阅读
大数据时代网络安全形势日趋严峻,本文提出了一种基于深度学习的异常入侵检测模型。首先,将网络流量数据进行数据预处理:针对网络流量数据的高维特征影响检测效率的问题,使用PCA等方法进行数据降维;其次,使用深度神经网络方法对预处理后的网络流量数据进行训练和类别预测;最后,使用混淆矩阵对模型输出结果进行评估,并与KNN和SVM两种经典算法进行对比。经过实验对比,本文模型均优于KNN算法和SVM算法,在准确率、召回率、F1-Score方面相比KNN和SVM的检测率提高2%。因此,本文模型有效提高了异常入侵检测的检测率,加强了网络安全。
In the era of big data network security situation is more and more serious, this paper presents a deep learning based on the intrusion detection model. Firstly, the network traffic data is preprocessed by data: aiming at the problem that the high dimensional features of network traffic data affect the detection efficiency, the data dimensionality reduction is carried out by using PCA and other methods. Secondly, the neural network method is used to preprocessed network traffic data Training and category prediction. Finally, the confusion matrix is used to evaluate the output of the model and compared with two classic algorithms, KNN and SVM. Through experimental comparison, the proposed model is superior to KNN and SVM algorithms, and the detection rate of KNN and SVM is improved by 2% in terms of accuracy, recall and F1-Score. Therefore, this model effectively improves the detection rate of anomaly intrusion detection and enhances the network security.