Statistical Learning for Technical Portfolio Management

ApSTAT Technologies Inc. provides insurance companies with analytical systems based on exclusive data mining technologies, enabling insurers to maximize the profitability of their operations. The company is working with a large financial institution to develop new financial products based on statistical learning techniques. The research team based at l’Université de Montréal had two objectives. The first was to develop a modular framework which would combine a large class of time-series processing modules in combinatorial fashions in a way that would make them accessible to an econometrician with little knowledge of computer science principles. Secondarily, the team was to improve the performance of several risk management models using state-of-the-art machine learning techniques. In the end, the team managed to improve the readability and usability of the company’s time-series processing tools as well as augment the robustness of the financial models used to build trading portfolios.

Intern: 
Christian Dorion
Faculty Supervisor: 
Dr. Yoshua Bengio
Province: 
Quebec
Discipline: 
Program: