Temporal Pattern of Human Post and Interaction Behavior in Qzone

摘要

The quantitative analysis of human pattern is an effective way to understand the complex social system. Relevant empirical studies reveal that human behavior in time follow a power-law distribution rather than a Poisson distribution. This paper aims at conducting statistical analyses based on the records of Qzone posts and interactive messages. The results show that the inter-event time distribution of posts and interaction follows a power-law distribution. Additionally, the time intervals of post comments differ from the time intervals of interact and in that there exists a clear cut-off point in the distribution of posting time, which indicates the subjection to a two-stage power-law. At the individual level, the posting time distribution exhibits fat tails. The analysis of post behavior indicates that there is a monotonous and negative relationship between the activity level and power-law exponent. The characteristics of local peaks also illustrate the burstiness and memory of post behavior.