Supervisor Recommendation in a Research Social Network


Student thesis: Doctoral Thesis

View graph of relations


  • Mingyu ZHANG

Related Research Unit(s)


Awarding Institution
Award date5 Jul 2016


Supervisor selection is important for students in their future studies and research careers. Currently, students rely on information search or friends’ recommendations to find potential research supervisors. However, due to problems of incomplete and asymmetric information between students and supervisors, it is not so easy for students to find suitable supervisors that match their research interests as well as their personality styles. There is a pressing need to provide an intelligent supervisor recommendation system to assist students to make wise decisions.

Several methods are proposed to solve the problems of supervisor selections in existing studies. They mainly consider topic-relevance information and candidate’s quality, but ignore the significance of connectivity dimension. Individual-level connection can be extracted from friend networks on research social platforms. Institutional-level connection can be mined through collaboration networks of published papers. In addition, current methods focus on meeting one-sided requirements of personality styles, but they neglect the personality matching between students and supervisors.

This thesis proposes a novel supervisor recommendation method that integrates relevance, connectivity, quality and personality dimensions in a research social network platform. The relevance dimension measures the similarity degree between a target student and potential supervisors in academic information aspects. This dimension is computed by a method of discipline-supervised semantic keyword matching. The connectivity dimension combines the individual-level connection and the institutional-¬level connection. In contrast to the traditional assessments of a researcher’s quality, the researcher’s social popularity on a research social network is complementally considered in this work. Subsequently, a personality-matching aided dimension is employed to analyze the matching degree of personality styles between students and supervisors. The proposed method is evaluated through a user study on a research social network. Average rate and normalized discounted cumulative gain metrics are used to compare the results from different methods. Experimental results show that our proposed method outperforms the baseline methods.

The present solution has been implemented as a recommendation service on a research social network (i.e., ScholarMate,, which aims to connect people together to research and innovate smarter. The supervisor recommendation method assists students in the decision process of supervisor selections. The proposed supervisor recommendation method is a novel way to provide personalized dynamic knowledge resources for students, which can also enhance their potential opportunities for social learning.