| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 5126790 | Social Networks | 2017 | 12 Pages |
â¢A methodology to detect the modularity structure of an evolving weighted directed network is proposed.â¢The method is based on tensor factorization and is applied to the Consolidated Banking Statistic.â¢Three communities arise according to the different banking systems.â¢The temporal activity of each community varies as events unfolded along the years.
The paper presents a new methodology aimed at detecting the modularity structure of an evolving weighted directed network, identifying communities and central nodes inside each of them, and tracking their common activity over time. The method is based on tensor factorization and it is applied to the Consolidated Banking Statistic, provided by the Bank of International Settlements. Findings show that data are well represented by three communities. The temporal pattern of each community varies according to the events involving the member nodes, showing a decrease of activities during crisis periods, such as the 2008 financial crisis and the European sovereign debt crisis.
