Seminars

Research and development - Seminars

Detecting anomalous payments networks:

A dimensionality reduction approach

This podcast explores the topology of financial connections, focusing on how interrelationships between financial institutions can amplify systemic risks. It highlights the rarity of reciprocal and transitive connections, which limits feedback effects within the financial system. Using empirical data, the temporal evolution of these networks is analyzed, highlighting the importance of regulation and the stability of these interconnections. In particular, we study the certificate of deposit networks, which present a density of approximately 4%, indicating that financial institutions do not tend to have frequent positions in assets issued by other entities, while the equity network shows virtually no reciprocity due to regulatory restrictions on cross-ownership of shares. This analysis highlights the low density and reciprocity, common features in financial networks according to the literature. In addition, the application of biological and neuroscientific concepts to financial research is discussed, highlighting how the principles of complex networks can anticipate crises, although in the Colombian context the lack of historical data complicates this analysis. Consequently, the importance of constantly monitoring financial networks to identify early signs of crises is emphasized.

Details:

Exhibitor:

Carlos Eduardo León Rincón

Date:

November 05, 2020

Play Video

Detecting anomalous payments networks

YouTube – Quantil Matemáticas Aplicadas

Attachments

Newsletter

Get information about Data Science, Artificial Intelligence, Machine Learning and more.