A Web-Based Network Traffic Prediction and Classification Application using Machine Learning

Web browsing has become a very common and almost indispensable activity for the ever-increasing number of internet users. However, with the increase in network traffic, the Quality of Service (QoS) of users is also impacted especially during peak utilization periods. It is therefore important to predict network traffic parameters such as bandwidth, and upload and download speeds which directly impact QoS. In this paper, a network analytics application is proposed whereby a browser extension is developed to analyse network traffic and perform prediction and classification. The extension sends requests to a Node.js server which provides real-time network traffic information to users and an indication of the QoS based on the parameters such as the latency, jitter, and upload and download speeds. The application can seamlessly be integrated in a Chrome browser and results show that it can effectively provide important network traffic data and classify the application type run in the browser.