Approximation of Missing Links in Stochastic Online Social Network Using Neural Network

In today’s vast, tightly connected world, social networks (SNs) are extremely important, having a deep impact on both our personal and professional lives. Link prediction in stochastic online social networks (SOSNs) is a challenging but important task that employs probabilistic models to estimate the potential links between network nodes. As stated earlier, because SOSNs bring unpredictability and uncertainty into network dynamics, link prediction is more challenging in SOSNs than in traditional SNs. On the other hand, it provides valuable insights into the connections and evolving hierarchies of these dynamic virtual communities. In SOSNs, a dynamic and complex virtual community paradigm, people’s behavior and link evolution are governed by probabilistic processes. Because SOSNs involve uncertainty and randomness, they present unique challenges for modeling and analysis compared to traditional online SNs. Nodes (individuals) in SOSNs employ stochastic actions to publish content, form new connections, and participate in debates. Due to the irrational interactions that cause nodes to make and break connections, the topology of the network is also prone to randomization. Therefore, to predict link formation and understand information dispersion in SOSNs, advanced modeling techniques such as stochastic processes and dynamic graph models are required.