Career age-aware scientific collaborator recommendation in scholarly big data

Abstract

Seeking a collaborator is one of the important academic activities of scholars because the right collaborators will help improve the quality of scholars' research and accelerate their research process. Therefore, it is becoming more and more important to recommend scientific collaborators based on big scholarly data. However, previous works mainly consider the research topic as the key academic factor, whereas many scholars' demographic characteristics such as career age, gender, etc are overlooked. It has been studied that scientific collaboration patterns may vary with scholars' career ages. It is not surprising that scholars at different career ages may have different collaboration strategies. To this end, we aim to design a scientific collaboration recommendation model that is sensitive to scholars' career age. For this purpose, we design a career age-aware scientific collaboration model. The model is mainly consisted of three parts, including authorship extraction from the digital libraries, topic extraction based on publication titles/abstract, and career age-aware random walk for measuring scholar similarity. Experimental results on two real-world datasets demonstrate that our proposed model can achieve the best performance by comparison with six baseline methods in terms of precision and recall.