DINGA: Discovery of Important Nodes in Social Networks Using Genetic Algorithms

Nowadays, the discovery of important nodes is among the main concerns of social networks analysis. Discovering important nodes has been done for wide range of purposes; from discovering leaders in terroristic networks to discovering influential people in advertisement networks. So far different criteria have been introduced for discovery of important nodes. Due to the diversity of graph structures from on side, and generality of previous methods in other side the accuracy of previous methods in discovering the important nodes is very limited. In this paper, a system called DINGA for discovering of important nodes in social networks with an unknown structure has been proposed. Our proposed system discovers the important nodes in social networks by employing a weighted combination of eight informative criteria of important nodes using genetic algorithm. The efficiency of the proposed solution has been investigated through simulation on the Enron e-mail network. Our proposed method outperforms the previous methods, in addition to random weight learner, 21% with respect to accuracy.