Spatial Structure,Hierarchy and Formation Mechanisms of Scientific Collaboration Networks:Evidence of the Belt and Road Regions

Scientific collaboration has become an important part of the people-to-people exchanges in the Belt and Road initiative,and remarkable progress has been made since 2013.Taking the 65 countries along the Belt and Road(BRI countries)as the research areas and using collaborated Web of Science(WOS)core collection papers to construct an international scientific collaboration matrix,the paper explores the spatial structure,hierarchy and formation mechanisms of scientific collaboration networks of 65 countries along the Belt and Road.The results show that:1)Beyond the Belt and Road regions(BRI regions),Central&Eastern Europe,China and West Asia&North Africa have formed a situation in which they all have the most external links with other countries beyond BRI regions.China has the dominant role over other BRI countries in generating scientific links.The overall spatial structure has changed to a skeleton structure consisting of many dense regions,such as Europe,North America,East Asia and Oceania.2)Within the Belt and Road regions,Central&Eastern Europe has become the largest collaboration partner with other sub-regions in BRI countries.The spatial structure of scientific collaboration networks has transformed from the‘dual core’composed of China and the Central&Eastern Europe region,to the‘multi-polarization’composed of‘one zone and multi-points’.3)The hierarchical structure of scientific collaboration networks presents a typical‘core-periphery’structure,and changes from‘single core’to‘double cores’.4)Among the formation mechanisms of scientific collaboration networks,scientific research strength and social proximity play the most important roles,while geographical distance gradually weakens the hindrance to scientific collaboration.

Measuring author influence in scientific collaboration networks

Purpose:The purpose of this paper is to measure author influence in scientific collaboration networks by considering the combined effects of multiple indexes.In the meanwhile,we intend to explore a method to avoid assigning subjective weights.Design/methodology/approach:We applied four centrality measures(degree centrality,betweenness centrality,closeness centrality and eigenvector centrality)and authors’published papers to the scientific collaboration network.The grey relational analysis(GRA)method based on information entropy was used to measure an author’s impact in the collaboration network.The weight of each evaluation index was determined based on information entropy.The ACM SIGKDD collaboration network was selected as an example to demonstrate the practicality and effectiveness of our method.Findings:Author influence was not always positively correlated with evaluation indexes such as degree centrality and betweenness centrality.This implies that combined effects of multiple indexes should be considered in author impact analysis.The introduction of the GRA method based on information entropy can reduce the interference of human factors in the evaluation process.Research limitations:We only analyzed author influence from the perspective of scientific collaboration,but the impact of citation on author influence was ignored.Practical implications:The proposed method can be also applied to detect influential authors in bibliographic co-citation network,author co-citation network,bibliographic coupling network or author coupling network.It would help facilitate scientific collaboration and enhance scholarly communication.Originality/value:This paper proposes an analytical method of evaluating author influence in scientific collaboration networks,in which combined effects of multiple indexes are considered and the interference of human factors is reduced in the evaluation process.

Clustering approach based on hierarchical expansion for community detection of scientific collaboration network

This study presents a clustering algorithm based on hierarchical expansion to solve the problem of community detection in scientific collaboration network. The characteristics of achievements information related to scientific and technological domains are analyzed,and then an ontology that represents their latent collaborative relations is built to detect clusters from the collaboration network. A case study is conducted to collect a data set of research achievements in the electric vehicle field and better clustering results are obtained. A hierarchical recommendation framework that enriches the domain ontologies and retrieves more relevant information resources is proposed in the last part of this paper. This work also lays out a novel insight into the exploitation of scientific collaboration network to better classify achievements information.

Step-by-Step Random Walk Network with Power-Law Clique-Degree Distribution

We propose a simple mechanism for generating scale-free networks with degree exponent γ=3, where the new node is connected to the existing nodes by step-by-step random walk. It is found that the clique-degree distribution based on our model obeys a power-law form, which is in agreement with the recently empirical evidences. In addition, our model displays the small-world effect and the hierarchical structure.

Evolution model for scientific collaboration network with local-world information

In order to reveal the intrinsic properties of scientific collaboration networks, a new local-world evolution model on a scientific collaboration network is proposed by analysing the network growth mechanism. The act degree as the measurement of preferential attachment is taken, and the local-world information of nodes is taken into account. Analysis and simulation show that the node degree and the node strength obey the power-law distribution. Low average path length and high clustering coefficient are approved. Experiment indicates that the model can depict efficiently the topological structure and statistical characteristics of real-life scientific collaboration networks.