Link Prediction in Social Networks Using Node Embeddings with Nature Inspired Approach

Social networks are important for tasks such as transmitting information, predicting the outcomes of an event, and predicting new relationships. The link prediction task involves the prediction of new relationships that may be created over time. More formally, suppose we have a social network’s snapshot at the current time, we are interested in accurately predicting the edges that are not currently present but will be present after some time has elapsed. A new approach, Nature Inspired using CC-CD, has been proposed in this paper which makes use of node embeddings and closeness centrality. We have used the cosine distances of node embeddings as a proxy of distances and Closeness Centrality as a proxy of masses in Newton’s Gravitational Law for prediction of new links.