Finding a team of experts in social networks
Proceedings of the 15th ACM SIGKDD international conference on Knowledge discovery and data mining
Power-Law Distributions in Empirical Data
SIAM Review
Collaboration in computer science: A network science approach
Journal of the American Society for Information Science and Technology
| Hi-index | 0.00 |
Team science is a collaborative approach to research, typically with researchers drawn from different disciplines. Team science networks have certain unique characteristics in their conception and intent that set them apart from other commonly studied social and collaboration networks. We study the structural properties, and present metrics for collaborative performance assessment in two real-world team science networks initiated by the Army Research Lab. We model a team using a higher-order generalization of an edge called a simplex. A simplex captures group relationships distinct from the union of pairwise relationships. Our evaluation using a rigorous methodology reveals that the distributions of vertex and facet degrees (the number of maximal groups that a vertex belongs to) follow a power law, but with exponential cut-off at the tail in most cases. We propose metrics for quantitatively assessing the extent of intra-team and extra-team collaborations, and compare their effectiveness vis-a-vis our intuitive notions. Our work can be used as the basis for generative models, and for evaluating the collaborative performance of team science networks.