Improving QS Rank Based on The Classification of Authors Research Collaboration Using Machine Learning Techniques

Abstract

The importance of universities' global ranking lies in providing a trusty resource, which helps students in choosing the right place to complete their academic future. The global ranking systems are based on several metrics that focus on the study environment, the quality of the provided services, the scientific publications, and the extent of the authors' strength. Quacquarelli Symonds (QS) is the most popular global ranking system, it has Citations Per Faculty (CPF) evaluation metric, which constitutes 20% of the total ranking score. In this research, we aim to find the effect of the research collaboration on increasing the CPF score, in which we apply descriptive analytics on a dataset for Jordan University of Science and Technology (JUST) authors, that is scrapped from the official websites of Google Scholar and Researchgate. Then, we find the authors who have a moderate collaboration through building a classification model using machine learning techniques. The results proved that the research collaboration has a significant impact in increasing authors publications that positively correlated with their total citations, which in turn gives a great opportunity to increase the CPF score. Also, the Support Vector Machine classifier has obtained a 95.27% level of accuracy, which considers as an efficient method in classifying the authors research collaboration into strong and moderate collaboration. Finally, the proposed method can be used to improve the QS ranking and obtain a high scientific standing level for academic institutes.