Abstract:

We discuss a practical machine learning approach to construct portfolios with mean-reverting price dynamics. Our objectives are threefold: (1) design a portfolio that is well-represented by a mean-reverting process with parameters estimated by maximum likelihood, (2) select portfolios with desirable characteristics, such as high mean reversion, and (3) build a parsimonious portfolio, i.e. find a small subset from a larger collection of assets for long/short positions. Our data-driven method combines statistical learning and optimization. We present a specialized projected gradient algorithm to solve the constrained non-convex problem embedded in the trading problem. Numerical examples using empirical price data are provided.

Biography:

Tim Leung is the Boeing Endowed Chair Professor in the Department of Applied Mathematics and the Director of the Computational Finance & Risk Management (CFRM) program and Quantitative Analytics Lab at University of Washington in Seattle. He has previously taught in the Department of Applied Mathematics & Statistics at Johns Hopkins University and in the Department of Industrial Engineering & Operations Research at Columbia University. He obtained his BS from Cornell University and PhD from Princeton University. His research in Quantitative Finance has been funded by the National Science Foundation (NSF). He has published over 70 peer-reviewed articles and several books on the topics of Mean Reversion Trading, ETFs, and more. Professor Leung is on the advisory board of the AI for Finance Institute and the editorial board of multiple journals. He has served as the Chair for the INFORMS Finance Section as well as the Vice Chair for the SIAM Activity Group on Financial Mathematics & Engineering.

Zoom Meeting：https://us02web.zoom.us/j/87874263432

Meeting ID：878 7426 3432

Password：150522