Interest-Based Mining and Modeling of Big Mobile Networks

Usage of mobile wireless Internet has grown very fast in recent years. This radical change in availability of Internet has led to communication of big amount of data over mobile networks raising new challenges and opportunities for modeling of mobile Internet characteristics. While the traditional approach toward network modeling suggests finding a generic traffic model for the whole network, in this paper, we show that this approach does not provide enough accuracy for big mobile networks. We show that user interest acquired based on accessed domains and visited locations has a significant effect on traffic characteristics of big mobile networks. Our case study based on a big dataset including billions of netflow record collected from a campus-wide wireless mobile network reveals the fact that domains and locations showing similar point of interests (e.g. domains of news agencies or locations of fraternities) mostly follow similar types of traffic distributions. For this purpose, we utilize a novel graph-based approach based on KS-test. We also show that interest-based modeling of big mobile networks based on visited domains and locations can significantly improve the accuracy and reduce the KS distances by factor of 5 comparing to the generic approach.