Navigating the Scholarly Networks: A Comparative Analysis of Link Prediction Techniques on Network of Research Articles

Abstract

In this work, different link prediction techniques in a network of research articles have been studied. The network of research articles can play a crucial role in the spread of scientific knowledge. The prediction of potential future links within this network holds significant implications for advancing research recommendations and academic collaborations. The link prediction techniques used in different complex networks like social networks recommend potential or useful links which, in this study, has found appropriate research articles. In this study, many traditional link prediction techniques have been applied to a network of research articles. The network has been developed by the authors by collecting real research articles from the chosen fields for their experimentation. The performance of these techniques has been assessed using various evaluation metrics, such as the area under the receiver operating characteristic curve (AUC-ROC), precision, and recall determining their effectiveness in predicting future links within the formed network of research articles. The findings have revealed the strengths and limitations of each link prediction technique, shedding light on their applicability to real-world networks. Further, deeper insights into the scalability of different methods through their efficiency in this domain have also been provided. © 2025 American Institute of Physics Inc.. All rights reserved.