A new approach for modeling researcher community based on scientific article metadata using natural language processing

摘要

The Ministry of Research, Technology, and Higher Education of Indonesia has requirements to answer five main questions, i.e. what are the most popular research topics; which research topics tend to diminish; what are the trending research topics; who are the most influential researchers in a specific research topic, and; who are leading researchers in a specific topic. This study is a part of our ongoing main project to develop a scientific community model based on scientific article publication. This paper introduces a new approach to model a virtual researcher network using metadata of articles. We modified PageRank to develop a term graph and author-term graph for this purpose. Based on the preliminary results, the model could answer the five aforementioned questions.