Deep Learning Based Network Traffic Analysis Using Modified Hybrid Methodology Comparing with SVM to Improve Accuracy

Users of networks are placing increased expectations on the speed and quality of the services provided by networks as a result of the fast advancement of network technology. As a result, one of the problems in the industry of network operation and maintenance management is to manage and regulate diverse network business traffic through efficient technological methods, differentiate between services, provide varied quality assurance, and fulfill the business demands of users. The identification of network traffic is a useful technological tool that may differentiate between the traffic generated by various applications. Through the processes of classifying, identifying, and distinguishing the application of network traffic, various types of traffic on the network may be provided with tailored network services, which in turn improves the quality of network services and the level of user satisfaction. The accurate identification of network traffic is not only a crucial foundation for the monitoring and data analysis of network traffic, but it is also the key to improving the overall quality of user service. Using a Hybrid Model that we've dubbed ANNSVM, the primary emphasis of this article is on doing an analysis of the data traffic on a 5G network. The acronym ANNSVM stands for Artificial Neural Network Support Vector Machine and combines the two terms. The term “artificial neural networks” (ANNs) refers to computer systems that utilize “learning algorithms,” which are programmes that can autonomously make modifications, or “learn,” when presented with new information. Because of this, they are a very useful tool for non-linear statistical data modeling, and SVM may function as a binary classifier. On the basis of the test data, the average classification accuracy is 98.8 percent, significantly exceeding other approaches that are already in use.