Machine learning and high value payments systems: uncovering patterns, hunting for anomalies.

2 Views
Published
THE 2021 PAYMENTS CANADA SUMMIT
The SUMMIT is Canada’s premier payments conference. Canada is undergoing a massive transformation in its payments ecosystem. The SUMMIT is the ideal meeting point for the payments community to exchange ideas on the future-state of the payments industry. Canada has a leading role to play in evolving the global payments ecosystem, and The SUMMIT is the place where we will see this transformation in action and the impact it will have on us as a nation, as an industry, and as individuals. Learn more at thesummit.ca.

SPEAKERS
Segun Bewaji - Senior Economist - Payments Canada
Simone Giansante - Assistant Professor, Finance - University Of Bath, School of Management
Anne Butler - Chief Legal Officer & VP, Research & Policy - Payments Canada

SESSION DESCRIPTION
This two-part session will discuss: De-Tiering Large Value Payment Systems. The Case of Canadian LVTS

A sampling approach to the problem of de-tiering payment systems that are characterized by few direct members. We apply physical reasoning to directly estimate all payments between any initial payer and final payee from payment data of direct members of a tiered payment system. We employ this methodology to the Canadian LVTS using proprietary data of payments as well as tiering structure from the direct members.

Detecting outliers in large value payment systems using neural networks: Does memory improve performance? Large value payment systems are a critical component of a country’s financial system and economy, supporting billions of dollars in real and financial transaction activity each day. These systems must comply with international standards on such things as liquidity risk management, and are generally subject to strong regulatory and operational oversight.

Our presentation will explore how machine learning can improve the detection of liquidity stresses in the Canadian and Dutch large value payment systems, with a focus on how adding an additional memory layer can improve performance.
Category
Management
Be the first to comment