Measuring the social influences of scientist groups based on multiple types of collaboration relations

Scientists often collaborate with each other and may produce social influences through their collaboration on scientific activities. While the subject of ranking scientists has received significant attention in previous studies, these studies often examine the social influences of individual scientists and are based on the assumption that all collaborations between scientists are of the same type. However, these two limitations do not match real scientific collaborations. Currently, scientists are often associated with groups in which the scientists always study related research topics or have close collaboration relationships. Moreover, current scientists often collaborate through multiple relationship types, and different types of relationships may have different effects on the social influences of the scientists. To solve these two problems, this paper presents a model that measures the social influences of scientist groups based on multiple types of collaboration relationships. The model addresses two general group types (hierarchical and nonhierarchical) and two general collaboration situations: one is that the multiplex collaboration relationships are independent, and the other is that the multiplex collaboration relationships are correlated. The model mainly adopts the linear weighted sum methodology, which can make the time complexities of key algorithms low. Finally, we create some case studies and make experiments to demonstrate the effectiveness of our model.