A method for classify network traffic with optimal individual convergence rate

Abstract

A method for classify network traffic with optimal individual convergence rate is disclosed, which is a novel stochastic one-step ladder algorithm and can solve that problem of large-scale network data classification. The method includes following steps of inputting network traffic, carry out the necessary pre-processing work, divide training samples and test samples; adopting Only one training sample randomly selected for each iteration, and a classification algorithm with optimal individual convergence rate for the training model. Calculating The sparse weight w of each iteration, and the objective function value and individual convergence rate c according to w. Finally, the model is tested with test samples, and the accuracy of network traffic classification is obtained. The invention provides a network traffic classification method, which has an individual optimal convergence rate and can effectively solve the problems of large-scale network traffic classification and identification. No specific protocol parsing is needed and the characteristics of classification can be reflected in real-time, so the method adopted has certain universality.

Description

technical field [0001] The invention relates to the field of data classification and identification, in particular to a network traffic classification method with an optimal individual convergence rate. Background technique [0002] As we all know, today, with more and more emphasis on personal privacy protection and network security, large-scale network traffic brings many new challenges to network security. At present, the state has established a working group on network security and informatization, focusing on solving security and developmental issues in the network. The current network security situation is complex and severe. Effective identification and management of network traffic is of great significance to protecting user information, supervising illegal data, detecting network attacks, and maintaining network security. At present, traditional classification methods in network traffic classification research are gradually being replaced by machine learning method...

AI Technical Summary

Problems solved by technology

[0003] The current network traffic data is generally characterized by many samples, high dimensionality, and sparse redundancy. The traditional batch processing mode is dif

Images