A Graph Based-Novel Framework for Social Synchrony Detection Using Influential User and Event Detection Approach

Social synchrony (SS) is the latest dynamic and intricate phenomenon witnessed on popular online social networks such as Twitter, Facebook, etc. A large population of users acting in synergy within a confined time constitutes social synchrony. The paper proposes a novel framework, Probability-based Influential User and Event Detection for Social Synchrony (PIE-SS), offering insights into social synchrony on the popular social network Twitter. The purpose of this study is to consider not only hashtags but also all user interactions, such as retweets, replies, mentions, and hashtags to offer a comprehensive view of the Twitter network. The study presents a probabilistic model in collaboration with temporal graphs, wherein the dynamic Twitter network has been divided into static sub-graphs across time, and the probability function monitors each sub-graph user states. The model considers the state of the nodes at each time frame and accordingly evolves the sub-graphs at respective time steps. Consequently, modeling temporal dependencies in graphs offers a better understanding and representation of the network. Compared to several state-of-the-art technologies mentioned in the existing literature, the proposed model’s accuracy, precision, recall, and F1-score values are reported as 0.899, 0.902, 0.896, and 0.914, respectively which beyond the baseline of similarly domain technologies as well as provide evidence of the effectiveness of the suggested PIE-SS strategy.