Algorithm for Recognition of Network Traffic Anomalies Based on Artificial Intelligence

Abnormalities in network traffic can be caused by malfunctioning network equipment, accidental or intentional actions by users, or the actions of attackers. Thus, for reliable data transmission in the information network, it is necessary to take measures to detect anomalies in a timely manner and take measures to eliminate them. Therefore, in order to ensure reliable data transmission in the network, the development of new methods for detecting anomalies is of urgent importance. This work is devoted to the development of an improved algorithm for recognizing network traffic anomalies based on artificial intelligence. On the basis of the conducted analysis and research, an improved algorithm was developed for the most accurate determination of an abnormal state. The principle component analysis algorithm was taken as a basis and a type of Generative adversarial network algorithm, a machine learning algorithm without a teacher, was added to it, namely BIGAN, which uses an encoder in its activity, namely, thanks to its E encoder, it is able to detect anomalies in input and processed data, which made it possible to detect network traffic anomalies with greater accuracy and in less time.