Detecting Hotspots in Interdisciplinary Research Based on Overlapping Community Detection

Abstract

Disciplinary fusion has been observed in a range of previous studies, which creates great benefit in science, technology, and innovation. However, how to detect and distinguish the hotspots in interdisciplinary is a challenge for not only researchers but also stakeholders in government and industry sectors. A keywords' co-occurrence network is constructed by using academic articles published between 2012 and 2017 in the field of Information Science and Artificial Intelligence. Then statistical methods are applied for finding the regularities of distributions of term frequency and keywords' k-cliques. Furthermore, the software CFinder, which is based on clique percolation method (CPM) algorithm and is used for detecting overlapping communities, is utilized to visualize the network. At last, we find that "Social media", "conceptual model, "Big Data" and "Crowdsourcing" are the hotspots of interdisciplinary research in this case.