147614 - Link prediction of the world container.pdf (2.86 MB)
Link prediction of the world container shipping network: A network structure perspective
Although the world container shipping network (WCSN) has gradually been shaped with ever-increasing complexity in link evolution over the last decades, its evolving mechanism remains to be unveiled. This motivates us to explore the evolutionary pattern of the WCSN, which can be achieved by advancing the existing link prediction models. Using the k-shell decomposition method, the network hierarchy can be decomposed and evaluated by four indices which are KS-Salton, KS-AA, KS-RA, and KS-LRW. The results show that the network hierarchy depends largely on trade patterns and demonstrates certain geographic characteristics. Meanwhile, the KS-LRW index performs best and, therefore, is further simulated for the future WCSN by predicting its top 1677 potential edges, which significantly enhances the overall network connectivity and efficiency. These findings create profound implications for shipping companies to strategically reduce the trail cost for new lines by analyzing the network data.
History
Publication title
ChaosVolume
31Pagination
1-20ISSN
1089-7682Department/School
Australian Maritime CollegePublisher
A I P Publishing LLCPlace of publication
United StatesRights statement
© 2021 Author(s). This article may be downloaded for personal use only. Any other use requires prior permission of the author and AIP Publishing. This article appeared in Jiawei Ge, Xuefeng Wang, and Wenming Shi, Link prediction of the world container shipping network: A network structure perspective, Chaos 31, 113123 (2021) and may be found at https://doi.org/10.1063/5.0056864Repository Status
- Restricted