Predicting social ties in mobile phone networks

A social network dynamically changes since the social relationships (social ties) change over time. The evolution of a social network mainly depends on the evolution of the social relationships. The social-tie strengths of person-to-person are different one another even though they are in the same group. In this paper we investigate the evolution of person-to-person social relationships, quantify and predict social tie strengths based on call-detail records of mobile phones. We propose an affinity model for quantifying social-tie strengths in which a reciprocity index is integrated to measure the level of reciprocity between users and their communication partners. Since human social relationships change over time, we map the call-log data to time series of the social-tie strengths by the affinity model. Then we use ARIMA model to predict social-tie strengths. For validation of our results, we used actual call logs of 81 users collected for a period of 8 months at MIT by the Reality Mining Project group and also used call logs of 20 users collected for a period of 6 months by UNT's Network Security team. These users have around 5000 communication partners. The experimental results show that our model is effective. We achieve prediction performance with accuracy of average 95.2% for socially close and near members. Among other applications, this work is useful for homeland security, detection of unwanted calls (e.g., spam), and marketing.