Community Aliveness: Discovering Interaction Decay Patterns in Online Social Communities (1707.04477v1)

Abstract: Online Social Communities (OSCs) provide a medium for connecting people, sharing news, eliciting information, and finding jobs, among others. The dynamics of the interaction among the members of OSCs is not always growth dynamics. Instead, a $\textit{decay}$ or $\textit{inactivity}$ dynamics often happens, which makes an OSC obsolete. Understanding the behavior and the characteristics of the members of an inactive community help to sustain the growth dynamics of these communities and, possibly, prevents them from being out of service. In this work, we provide two prediction models for predicting the interaction decay of community members, namely: a Simple Threshold Model (STM) and a supervised machine learning classification framework. We conducted evaluation experiments for our prediction models supported by a $\textit{ground truth}$ of decayed communities extracted from the StackExchange platform. The results of the experiments revealed that it is possible, with satisfactory prediction performance in terms of the F1-score and the accuracy, to predict the decay of the activity of the members of these communities using network-based attributes and network-exogenous attributes of the members. The upper bound of the prediction performance of the methods we used is $0.91$ and $0.83$ for the F1-score and the accuracy, respectively. These results indicate that network-based attributes are correlated with the activity of the members and that we can find decay patterns in terms of these attributes. The results also showed that the structure of the decayed communities can be used to support the alive communities by discovering inactive members.