Dynamic traffic classification algorithm and simulation of energy Internet of things based on machine learning

With the rapid development of information technology, a large amount of traffic generated by various Internet applications occupies a large amount of network resources. It poses a huge challenge to service quality and has a negative impact on Internet security. In order to utilize network resources effectively and provide effective management and control measures for network administrators, network traffic classification technologies is a hot topic for scientists to identify application layer protocols. Today, there are more and more applications based on TCP/IP. With the emergence of various anti-surveillance applications, traditional port and application-based identification methods are difficult to meet current or future traffic identification requirements. It has become a very challenging problem to require more efficient, accurate, intelligent and real-time Internet traffic identification. The Internet of Things is a new network concept proposed by people who based on Internet prototypes. It enables the end user of the system can carry out communication and exchange of information and data between any project. In recent years, with the continuous advancement of Internet of Things technology, the coverage of the Internet of Things has become very wide, and the number of different types of networks that make up the Internet of Things is also increasing. This paper aims to find the dynamic network traffic classification problem of hybrid fixed in dynamic network and dynamic network in mobile network, and gives a reasonable mapping scheme. The dynamics of network traffic for Internet of Things are reflected fully and will not cause route flapping. The simulation results show that the decision tree classification algorithm in machine learning has higher efficiency, and improves the utilization of network resources.