基于融合注意力机制的并列GRU应用层协议识别方法

Parallel GRU Application Layer Protocol Identification Method Based on Fusion Attention Mechanism

针对并列门控循环神经网络(GRU)算法在处理序列信息时无法直接交流和共享信息,导致模块之间信息流动受限,从而影响准确率的问题,提出了一种基于融合注意力机制的并列GRU应用层协议识别方法。该方法利用注意力机制获得并列GRU算法中不同时间点输出之间的重要关系,使得模型能更好地捕获序列数据的特征信息,从而提高算法的准确率。在UNSW-NB15数据集上进行实验,结果表明:与并列GRU算法相比,所提算法的识别准确率提升了9.3%,且与其他代表性算法相比,准确率均有所提高。

In order to solve the problem that the parallel GRU(gated recurrent unit)algorithm cannot directly communicate and share information in the processing sequence information,resulting in the limited information flow between modules,thus affecting the accuracy,a parallel GRU application layer protocol recognition method based on fusion attention mechanism was proposed.In this method,the attention mechanism is fused on the basis of parallel GRU,and the important relationship between the output of different time points in the parallel GRU algorithm is obtained by using the attention mechanism,so that the model can better capture the important relationship between the feature information of the sequence data and the time points,so as to improve the accuracy of the algorithm.The experimental results show that the accuracy of the proposed method is improved by 9.3%compared with the parallel GRU algorithm,and the accuracy is improved compared with common algorithms.