Internet of Things dynamic flow classification method based on machine learning

Abstract

The invention discloses an Internet of Things dynamic flow classification method based on machine learning, and the method comprises the following steps: 1), collecting Internet of Things dynamic flowdata which is a protection message set passing through a network observation point in a fixed time interval; 2) performing network traffic feature extraction on the Internet of Things dynamic trafficdata acquired in the step 1); 3) according to the network flow characteristics extracted in the step 2), performing network flow matching classification by using a machine learning method to completemachine learning-based Internet of Things dynamic flow classification. According to the invention, the method can efficiently, accurately and real-timely realize Internet of Things dynamic flow classification.

Description

technical field [0001] The invention relates to a method for classifying dynamic traffic of the Internet of Things, in particular to a method for classifying dynamic traffic of the Internet of Things based on machine learning. Background technique [0002] The popularity of the Internet has brought convenience to users, and network traffic has also surged. How to effectively manage the network, control traffic, rationally allocate and utilize network resources, and guide user resource analysis has become a key issue in the field of network supervision. Accurate network traffic classification is an important basis for network traffic monitoring and network traffic analysis. [0003] The current status of domestic network traffic management is that most network management departments have not established a system that can fully meet the above-mentioned network management requirements and traffic distribution. [0004] Most network management systems only use some common netw...

AI Technical Summary

Problems solved by technology

Cons: Both network traffic analysis systems have significant technical limitations and do not provide sufficient network traffic information

[0005] The traffic classification method based on the support vector machine uses too many data flow features, and the defect: it leads to an increase in computational complexity, resulting in an excessive amount of calculation in practical applications

However, with the continuous advancement of network technology, many network applications in the market have begun to use dynamic ports, which will lead to a decrease in the accuracy of method classification

[0007] Therefore, traditional port and application-based payload identification methods are difficult to meet current or future traffic identification, requiring more efficient, accurate, and real-time Internet traffic identification to become a very challenging problem

Images