Strategic hedging and portfolio allocation with decision trees

In this project, we address the problem of designing artificial learning-based methodology adapted to currency exchange risk valuation and classification. The goal is then twofold: 1) design and implement client risk classification methodology based on currency risk exposure and, 2) study the problem of dynamical optimal allocation within currency-risk hedging portfolios using non-parametric forecasting methods. In terms of risk valuation we seek to implement state-of-the-art risk classification techniques that would allow our partner to improve the robustness of their strategies overall performance. In terms of optimal allocation in their hedging portfolios, the firm looks for alternative models that would allow them to forecast a set of currency and currency related indices and derivatives within their hedging allocation scenario-based methodology.

Faculty Supervisor:

Mathieu Boudreault;Manuel Morales

Student:

Partner:

FinMetrix Inc

Discipline:

Mathematics

Sector:

Professional, scientific and technical services

University:

Université du Québec à Montréal

Program: