Identification of tidal-traffic patterns in metro-area mobile networks via Matrix Factorization based model

Due to the highly predictable daily movements of citizens in urban areas, mobile traffic shows repetitive patterns with spatio-temporal variations. This phenomenon is known as Tidal Effect analogy to the rise and fall of the sea levels. Recognizing and defining traffic load patterns at the base station thus plays a vital role in traffic engineering, network design and load balancing since it represents an important solution for the Internet Service Providers (ISPs) that face network congestion problems or over-provisioning of the link capacity. Previous works have dealt with the classification and identification of patterns through the use of techniques, which inspect the flow of data of a particular application. But they assume prior knowledge on the stream of data packets, making the trend identification much inefficient. Recent methods based on machine learning techniques build their classification models based on sample data collected at certain points of the network with high accuracy. Therefore, in this paper, we address the problem by applying matrix factorization based models on real-world datasets, identifying typical patterns from data streams, which frequently occur in the network, without investigating the type of flows. For that, we propose a Collective Non-negative Matrix Factorization based model combining multi-source data, such as point of interests attributes, traffic data and base station information, identifying the basic patterns of those areas of the city that present the same type of attributes. The experimental results show the effectiveness of our proposed approach compared with the baselines.