Heterogenous Graph Mining for Measuring the Impact of Research Institutions

Mining influential nodes in a social network for identifying patterns or maximizing information diffusion has been an active research area with many practical applications. In the research community, influential institutions usually attract denser attention than others. Based on the prediction on how many papers will be accepted by some top conferences held in 2016, the KDD Cup 2016 hosts an international competition for evaluating the importance of academic institutions. This paper describes our solution to the competition. Specifically, the proposed scheme involved in the competition mainly comprises of feature engineering and application of decision tree models. Finally, as claimed by the competition organizer, our approach scored 0.6599, 0.8169, 0.7213 with NDCG@20 in phases 1-3, and resulted in 0.7472 in overall score. With the above scores, our team ranked the first place in phase 2 and fourth place in overall rank.