Link Prediction in Stochastic Online Social Network Using Machine Learning Algorithm

Social networks (SNs) have a huge and diverse impact. Each and every persons are connected with SNs by directly or indirectly. They provide chances for community development, information sharing, and social interaction. But there are drawbacks as well, including as addiction, privacy issues, and the dissemination of false information. Predicting the possibility of a link existing between two nodes in a network is the goal of the network analysis task known as “link prediction.” Because graph neural networks (GNNs) can capture intricate relationships inside networks, they have become an effective tool for this endeavour. This research presents a new approach to link prediction based on optimal networks trained using link prediction models as output data and neural networks trained on scale-free networks as input data. A greedy link pruning method is used to address the impact of generalization of neural networks on experiments. The suggested global network structure reliability and network efficiency are taken into account as the goal to fully assess the link prediction performance and benefits of the neural network approach.