I enthusiastically welcome doctoral students to participate in my research.  I work with doctoral students in two institutions:  (1) the Rutgers Business School — Newark and New Brunswick , and (2) the Department of Computer Science at Royal Holloway College , University of London .

In the Rutgers Business School, I work with full-time doctoral students in several different majors, including accounting, accounting information systems, finance, information technology, and management science.  I also serve as faculty director of the school’s doctoral program, and in this role I would like to encourage potential applicants to consult the program’s extensive website, including the frequently asked questions and the online application.

In the Department of Computer Science at Royal Holloway, I am affiliated with the Computer Learning Research Centre, where doctoral students work on the theory of machine learning and its applications to a wide variety of fields.  Potential students may apply online for full-time or part-time study.

Some topics for research with doctoral students

For more than thirty-five years, I have worked on the foundations of probability, statistics, artificial intelligence, machine learning, accounting, and finance.  I continue to be fascinated by new methods for prediction and the weighing of evidence, their connections with older ideas about probability, and their implication for the philosophy of probability.  I am particularly interested in applications of game-theoretic probability, defensive forecasting, conformal prediction, and the Dempster-Shafer theory of evidence.

I enjoy working with students in accounting, finance, and marketing who can formulate their own substantive questions, to which machine-learning methods can be applied.  I am also interested in working with students who want to explore the implications of the game-theoretic formulation of the efficient market hypothesis.

I would also like to work with students with strong enough computational or mathematical skills that they can help

·          extend existing systems for defensive forecasting, conformal prediction, and Venn prediction,

·          develop the predictive method of causal analysis to compete with the more established counter-factual method, or

·          develop systems that guide Dempster-Shafer analyses of evidence using Cournotian judgements.

These topics would be appropriate for students in information technology or management science at Rutgers or in machine learning at Royal Holloway.

Recent and current work with doctoral students

Tanya Levin completed a doctoral dissertation in information technology at the Rutgers Business School under my supervision in May 2004.  Entitled Using the Aggregating Algorithm for Portfolio Selection, it studied new strategies for selecting portfolios in the stock market.  These strategies were inspired by two streams of previous work: (1) work on universalization of strategies for portfolio selection, which began with Thomas Cover’s work on constant rebalanced portfolios, published in 1991, and (2) Vladimir Vovk’s aggregating algorithm and his recent work on Markov switching strategies.  Vovk’s idea of a Markov switching strategy had not previously been applied to portfolio selection.  This dissertation developed and implemented some novel but simple Markov strategies that were remarkably effective.  Ms. Levin is now assistant professor of management science at Queen’s Business School in Kingston , Ontario .

Elaine Henry completed a doctoral dissertation in accounting at the Rutgers Business School under my supervision in May 2005.  Entitled Are Investors Influenced by How Earnings Press Releases are Written?, it showed that investors do pay attention to the tone as well as the substance of earnings press releases.  This was the first work in accounting to make use of machine-learning methods for analyzing text to predict stock market reactions.  Ms. Henry is now an assistant professor at the University of Miami .

In 2004, as a doctoral student in finance at the Rutgers Business School , Yudan Zheng co-authored with me a note entitled “The Principle of Coherence.” It explains the game-theoretic version of the principle of no arbitrage.  In October 2007, Ms. Zheng completed a dissertation with Professor Darius Palia.  She is now an assistant professor at Long Island University in Brooklyn .

Wei Wu, a doctoral student in finance at the Rutgers Business School , is currently working under my supervision on a dissertation that uses the game-theoretic framework for probability to analyze autocorrelations and lagged cross-autocorrelation of stock returns in the United States .  Preliminary results show that the highly significant autocorrelations for portfolios of small-cap stocks are accounted for by transaction costs of 150 basis points or less.

Sam Ring, another doctoral student in finance at the Rutgers Business School , is currently working with me and Professor C. F. Lee on a dissertation that compares credit pricing models with machine-learning methods such as support vector machines and defensive forecasting.