Malicious traffic detection method integrating deep neural network and hierarchical attention mechanism

Abstract

The invention provides a malicious traffic detection method integrating a deep neural network and a hierarchical attention mechanism, which comprises the following steps: S1, acquiring original traffic data, and storing the acquired original traffic data as traffic data in a recognizable file format; S2, performing feature conversion on the traffic data stored in the step S1; S3, performing data packet segmentation on the traffic data converted in the step S2 to obtain data packet segments; S4, capturing feature information between the data packet segments through a time sequence processing feature vector; S5, performing allocation to obtain an attention vector; S6, performing feature fusion on the traffic data; S7, performing linear transformation on the features fused in the step S6; andS8, classifying the traffic data. According to the invention, malicious traffic can be detected, and the performance is enhanced.

Description

technical field [0001] The invention relates to the technical field of malicious traffic detection, in particular to a malicious traffic detection method that integrates a deep neural network and a hierarchical attention mechanism. Background technique [0002] With the continuous development of computer network, it is constantly changing the way of people's life, study and work, but it is facing various security threats, and this threat is becoming more and more serious. This led to network security, which includes policies and practices for preventing and monitoring computer networks and the resources that can be accessed through the network for unauthorized access, misuse, modification or denial. Network security mainly includes the confidentiality, integrity and availability (Confidentiality Integrity Availability, CIA) of its bearer information. Any kind of activity that attempts to undermine the CIA or bypass established network security mechanisms can be considered a...

AI Technical Summary

Problems solved by technology

But because the machine learning algorithm can only be used as a classifier, it has many limitations, and when the intrusion becomes more and more complex and diverse

Images