Network threat detection system based on auto-encoder integration

Abstract

The invention discloses a network threat detection system based on auto-encoder integration, and the system comprises: a network data obtaining module which is used for obtaining a network flow data feature extraction module from the Internet or a local file, and is used for extracting a flow feature to generate a feature vector; a feature clustering module which is used for grouping the featuresaccording to the correlation; a threat detection module which is used for detecting traffic abnormality by using an integrated model based on an auto-encoder; and a threat judgment module which is used for judging the network threat. According to the system, an unsupervised deep learning algorithm is used, network threat detection is carried out by improving an existing auto-encoder algorithm, detection of unknown threats is achieved, the time complexity of neural network algorithms such as an auto-encoder is reduced while the detection accuracy of the model and the real-time performance of the model are improved, and the system is simple in implementation method and flexible in means and high in practicability, can effectively detect network threats, and is irrelevant to specific hardware.

Description

technical field [0001] The invention relates to the technical field of computer network security, in particular to a network threat detection method based on autoencoder integration. Background technique [0002] The rapid development of the Internet has brought convenience to people's lives and brought opportunities for the development of enterprises, but at the same time the Internet has also brought new threats to modern society. Network threats at various levels emerge in an endless stream, such as viruses, Trojan horses, and DDoS attacks, which seriously threaten the security and interests of individuals and enterprises. It is an important research topic in the field of network security to discover and successfully prevent the losses caused by network threats in time. [0003] Traditional network threat detection methods are mostly based on signatures. This type of algorithm has a good detection effect on known threats, but the detection effect on unknown threats is of...

AI Technical Summary

Problems solved by technology

[0003] Most traditional network threat detection methods are based on signatures. This type of algorithm has a good detection effect on known threats, but the detection effect on unknown threats is often unsatisfactory.

However, the update speed of network threats is extremely fast, how to quickly discover new network threats is the problem that network threat detection systems are now facing

Images