I spent several years in the financial services industry researching the applicability of machine learned methods for predictive modeling in financial markets. The trading strategies that emerged from this research are offered to investors by institutions. The track record since May 2002 is comparable to top institutional Commodity Trading Advisors (CTAs) with a winning monthly percentage of roughly 58% with roughly equal sized average winning and losing months. Is that better than random? Almost definately.
My research focuses on building predictive models based data from a variety of sources such as financial times series, economic networks, blogs, and sponsored search markets. More recently, I have been working with IBM Research on the question of "What Data Should You Keep?" This view considers data as both an asset as well as a potential liability thereby taking a risk-based approach to Data Governance.
Professor
Head of Information Systems
Stern School of Business
New York University
44 West Fourth Street
KMC 8-97
New York, NY 10012
phone: 212.998.0816
fax: 212.995.4228
email: vdhar@stern.nyu.edu