Variable social vector clocks for exploring user interactions in social communication networks

Social network communication analysis has drawn widespread attention in recent years. Vector clocks can be applied to capture the most recent communication with all other local peers in a social network. A modification of conventional social vector clocks has been previously proposed to deal with the issue of poor scalability without keeping whole temporal views. In this paper, our proposed framework maintains the low bound of how out-of-date each peer can be with respect to others, and also considers the shortest friendship separation to restrict how far information may be transmitted along time-respecting paths. To quantitatively analyse the influence of user interactions over different limitations of friendship distance, we also provide an adaptive incremental updating approach that can exactly recover the real situation in a specified upper bound of friendship distance. Experimental results also show that social vector clocks can be efficiently exploited to improve memory space requirements.