Abstract

Finding the right researcher for a joint research project is crucial in improving the outcome of the research. However, owing to the recent popularity of multidisciplinary research, it is difficult to find an appropriate researcher familiar with your field of study. In this study, we constructed a co-authorship network as a heterogeneous graph, using three bibliographic information: author, paper, and field of the paper. In addition, we extracted researchers who may coauthor in the future using a Relational Graph Convolutional Networks (R-GCN).