Stepping beyond your comfort zone: Diffusion-based network analytics for knowledge trajectory recommendation
Interest in tracing the research interests of scientific researchers is rising, and particularly that of predicting a researcher's knowledge trajectories beyond their current foci into potential inter-/cross-/multi-disciplinary interactions. Hence, in this study, we present a method of diffusion-based network analytics for knowledge trajectory recommendation. The method begins by constructing a heterogeneous bibliometric network consisting of a co-topic layer and a co-authorship layer. A novel link prediction approach with a diffusion strategy is then used to reflect real-world academic activity, such as knowledge sharing between co-authors or diffusing between similar research topics. This strategy differentiates the interactions occurring between homogeneous and heterogeneous nodes and weights the strengths of these interactions. Two sets of experiments - one with a local dataset and another with a global dataset - demonstrate that the proposed method is prior to selected baselines. In addition, to further examine the reliability of our method, we conducted a case study on recommending knowledge trajectories of selected information scientists and their research groups. The results demonstrate the empirical insights our method yields for individual researchers, communities, and research institutions in the information science discipline.