A approach for a new collaboration recommendation in co-authorship networks based on Global Cooperation and Global Correlation

Abstract

In this paper, we propose a new collaboration recommendation in co-authorship networks to assist researchers in specifying existing research collaborations and strengthening them in the future. It is based on Global Cooperation and Global Correlation to further improve the recommendation performance. Global Cooperation relies on the connection between authors and their common research works. Global Correlation is determined through a topic modeling method, namely Latent Dirichlet Allocation (LDA). The proposed system determines the outcome based on specified thresholds for the Global Cooperation and Global Correlation. It is experimentally validated on a dataset of co-authorship networks published in the “Biophysical Journal” from 2006 to 2017.