Assessing the stability of collaboration networks: A structural cohesion analysis perspective

Abstract

In collaboration networks, a stable structure can lead to trust and enhance group members’ ties, in turn reducing conflicts and promoting communication and cooperation. Therefore, network stability assessment, especially for collaboration networks, is essential for facilitating the achievement of group goals. However, most previous studies have considered only a fundamental understanding of network stability from the perspective of network connectivity or interpersonal relationships. Few studies have been conducted to reveal the influence of endogenous structural cohesion on network stability. In fact, greater structural cohesion indicates greater adaptability in uncertain environments. Thus, we propose evaluating the stability of collaboration networks from a structural cohesion perspective. Our study focuses on two dimensions of structural cohesion: core member identification and structural robustness measurements. Considering the unique structure of collaboration networks, a new algorithm, named the improved K-shell decomposition algorithm, is proposed to identify the core member set embedded in the innermost layer of a network. Compared with traditional identification algorithms, our algorithm can achieve a better trade-off between computational accuracy and computational complexity. Experimental results obtained on real-world networks verify the performance of our algorithm. In addition, it was found that the stability of collaboration networks can be effectively improved through targeted prevention efforts at the core members identified by our algorithm.