A Survey on Mobile traffic prediction for cellular network using deep learning

Abstract

The network traffic’s analysis and prediction has many applications in an understandable group of areas which have interesting and many important studies. Various kinds of experiments are done and concluded to find different types of problems in the current computer network and its application. The analysis and prediction mechanism in this network traffic is a proactive method to make sure that this approach has a reliable, secure and qualitative network communication. Different types of methods are developed and experimented to analyze the network traffic which includes data mining mechanisms. To achieve efficient and reliable performance, a number of intriguing variations of network analysis and prediction strategies are used. The aim of this paper is to provide an overview of numerous network analysis and traffic prediction techniques. Previous research is examined for its uniqueness and laws. In addition, numerous fields of network traffic analysis and prediction that have been completed have been summed. This study provides the systematic and thorough analysis of the most recent research and study efforts focusing on the machine learning (ML)-based performance enhancement of the wireless networks, taking into account all layers of the protocol stack: PHY, MAC, and network. To help the experts of non-machine learning to understand all the considered methods, the relevant study and research work submissions are discussed first, then by providing the required context on data-driven mechanisms and machine learning. Afterwards, a systematic analysis of works using machine learning (ML) to refine wireless networking parameter settings for better network quality-of-service (QoS) and quality-of-experience is discussed (QoE). Initially, these works are divided into three categories. They are: MAC analysis,radio analysis and network prediction methods, with subcategories under each. Lastly, there is a discussion of open issues and wider viewpoints.