Dynamic network embedding enhanced advisor–advisee relationship identification based on internet of scholars

Abstract

Advisor–advisee relationship is a special social relationship and interpersonal relationship. In the era of scholarly big data, mining and analyzing this kind of academic relationships is of great significance. Though many studies explore the advisor–advisee relationships based on real-world dataset, the scale of their dataset is relatively small. Based on the assumption that advisor–advisee relationships are hidden in collaboration networks, this paper proposes a novel method by performing dynamic network embedding on internet of scholars. Specifically, we consider various scholar attributes and dynamic network embedding-based scholar vector as the input of supervised machine learning methods for advisor–advisee relationship identification. Experimental results on the real-world dataset show that our proposed method can achieve the best performance compared with several state-of-the-art methods.