Association analysis based on mobile traffic flow for correlation mining of mobile apps

As the smartphone market increasingly gains its prosperity nowadays, Internet traffic generated by mobile terminals is also booming. The detailed traffic flow data is helpful to discover association relationship of applications and user behavior patterns. In this paper, we study the correlation of mobile apps based on a mobile network traffic dataset from a data mining perspective. This dataset is acquired by a traffic monitor which is deployed in the backbone network of a mobile network operator in a province of China. We analyze both the mobile applications and the websites which the mobile user has accessed. Then an association analysis method is conducted to extract rules which can reflect characteristics and preference of users when accessing the Internet by smartphones. By employing the Apriori algorithm on our mobile traffic flow data, we discover some association rules among several mobile apps and websites which can be helpful for understanding the mobile network user behavior.