一种基于深度学习的网络异常流量检测算法

A Network Anomaly Traffic Detection Method Based On Deep Learning

为解决当前海量数据和分类不均匀数据流量的检测问题,提出一种基于深度学习的异常流量检测算法。该算法将FCM算法和GRNN相结合,采用FCM算法对数据流量样本进行聚类,然后使用GRNN对距离FCM簇中心最近的样本点进行卷积训练并迭代更新,直到获得一个稳定的簇中心;引入MFOA对FCM-GRNN进行参数调优,利用MFOA的全局寻优特性和三维空间搜寻方法,迭代优化找到最优的Spread值;使用KDD CUP99数据集进行试验,得出所提算法的检测率为91.36%,误检率为1.154%,所提算法具有较好的异常流量检测能力。

In order to solve the current problem of massive data and classified uneven data traffic detection,an anomaly traffic detection algorithm based on deep learning is proposed.This algorithm combines FCM algorithm with GRNN,uses FCM algorithm to cluster data flow samples,and then uses GRNN algorithm to conduct convolution training on the sample points closest to the FCM cluster center and iterate and update until a stable cluster center is obtained;MFOA algorithm is introduced to optimize the parameters of FCM-GRNN method,and the optimal Spread value is found by iterative optimization using the global optimization characteristics of MFOA algorithm and three-dimensional space search method;Using the KDD CUP99 data set,the test results show that the detection rate of the proposed method is 91.36%,and the false detection rate is 1.154%.The proposed method has a good ability to detect abnormal flow.