An Approach to Enhance Academic Ranking Prediction with Augmented Social Perception Data

Abstract

Academic ranking prediction are indicators that have significant influences on the decision-making process of stakeholders of universities. In addition, we are in digital age with a pandemic situation, social media and technology have revolutionized the way scholars reach and disseminate academic outputs. Thus, the ranking consideration should be adjusted by augmented social perception data, e.g. Altmetrics. In this study, dataset of 1,752,494 research outputs from Altmetric.com and Scival.com which published between 2015–2020 are analyzed. This study assesses whether there are relationships between various scholarly output’s social perception data and citations. Moreover, various machine learning models are constructed to predict the citations. Results show weak to moderate positive correlation between social perception data and citation. We have found that the outperforming prediction model is Random Forest regression. The finding in our study suggested that social perception data should be considered to enhance academic ranking prediction in conjunction with related features.