FINDING INFLUENTIAL NODES IN SOCIAL NETWORKS

Abstract

We can get useful information about personal preferences, hobbies, and connections through social networks. This data could be useful in the development of recommender systems, the prediction of social influence-based outcomes, and the acquisition of knowledge. The most influential nodes in the network, also known as spreader nodes, are a strategic technique of optimizing and tracking the influence and transmission of certain information. Despite the existence of a variety of methods for identifying influential nodes in a network, recent research shows that ensuring all-round performance of selected nodes based on influence, spread, and reach is a difficult challenge. We developed a hybrid filter-based approach in which nodes are filtered based on different centrality measures and the top filtered nodes are elected as spreaders in our research. Our proposed work beats all other relevant research works when tested on a range of real-life networks across numerous judging parameters, thanks to its strategic teaming of selected spreaders and overall performance in network simulations. Another approach is also discussed where we take in the advantage of community detection and neighborhood distinctness to find out the seeds set of the social graph.