BRF: Network Traffic Analysis based on Bagging Ensemble with Random Forest Classifier
N. Deeban, P. Shyamala Bharathi · 2022
The fifth-generation (5G) network provides support for a wide variety of systems, including applications that demand the highest level of security as well as dependable communication. Because of recent advancements in smart devices, we are currently experiencing an explosion in the generation of data as well as a heterogeneity that necessitates the development of new network solutions for improved traffic analysis and comprehension. In order to automatically manage the massive amounts of data, these solutions need to be both intelligent and scalable. It is now feasible and easy to deploy machine learning (ML) to solve complex problems, and its effectiveness has been validated in a number of different domains. This is due to the progress that has been made in high-performance computing. It is anticipated that the deployment of 5G wireless communication systems will start in the year 2020. The traffic management for 5G networks will present significant technical challenges due to the introduction of new use cases, new technologies, and new network architectures. In this paper, we have concentrated on analyzing network data traffic for 5G Network using an ensemble method known as Bagging ensemble with Random Forest. Bootstrap sampling is utilized in the standard ensemble method known as “bagging.” The random forest technique is an improvement on the bagging method that can result in better variable selection. First, we will go over the concept of bagging, and then we will go over the enhancement that brought about random forest. The random forest algorithm is objective because it consists of multiple trees, each of which is trained on a different subset of the data. When you have both categorical and numerical features to work with, the random forest algorithm performs particularly well. On the basis of the test data, the classification accuracy is an average of 96 percent, significantly outperforming other methods.