Summary
Predicting internet traffic in ISP networks is of utmost priority to the network service providers and vital towards maintaining the required level of SLA for its residential and business customers. This research project will analyse a comprehensive set of traffic foecast algorithms that can predict network network traffic in ISP networks. In this regard, a suite of conventional statistical approaches and state-of-the-art machine learning and deep learning techniques will be analyzed and compared. Besides, our research will make necessary adjustment and propose novel techniques to make further improvement in the forecast algorithms.