More Complex More Productive: Characterizing Top Universities Based on Research Publications

Exploring new scientific concepts and imparting knowledge are important roles of universities. Up to now, most information management study on institutional research output focuses on the number and excellence of paper. This paper proposes a new characterization method from the perspective of output and complexity to extract academic information. Top-ranked universities are selected to identify different performance through research production and complexity. The production indicator of different universities is calculated based on the annual number of research paper produced in each university. The complexity indicator of different universities is obtained according to weighted revealed comparative advantage over different research disciplines. By using an unsupervised competitive learning algorithm that considers four indicators simultaneously, we construct a coherent framework to seize the nature of universities' research output. As a key finding, we discover that university research complexity has a positive relationship with research production and a different cluster of universities has a different rate of rising of the positive relationship.