Andreas S. Weigend

• PhD Physics, Stanford 1991
• Diplom Physics and Philosophy, Bonn 1986

Research Interests

• Application Areas
• Computational Finance and Financial Engineering
• Time Series Analysis and Prediction
• Asset Allocation and Risk Measurement
• Decision Making Under Uncertainty
• Modeling Users on the Web
• Document Analysis and Content Delivery
• Techniques
• Model Building and Evaluation
• Data Mining and Knowledge Discovery
• Machine Learning and Computational Intelligence
• Finite Mixture Models
• Hidden Markov Models
• Neural Networks and Graphical Models

Brief bio (text only, with links)

My research focuses on extracting knowledge from (possibly quite large) data sets. I develop and apply state-of-the-art methods from modern time series analysis, statistical artificial intelligence and neural networks, to problems in business, marketing and finance, as is part of the Knowledge Discovery, Data Mining and Modeling Group at NYU/Stern.
As an example of the research, a joint project with the NYU Salomon Center analyzes the complete set of 30 million transactions from 3 years of T-bond futures to understand trading styles. Another example is the work on hidden Markov experts for trading and risk management. An application to predicting the daily probability distribution of S&P500 is available as html , as postscript (uncompressed) and compressed (gzipped) and and as pdf (portable document format) .
Current projects include Nonlinear Prediction of Conditional Percentiles for Value-at-Risk (with Isaac Chang), Computing Portfolio Risk Using Gaussian Mixtures and Independent Component Analysis (with Elion Chin and Heinz Zimmermann), and Option Pricing and Implied Risk-Neutral Densities (with Chris Pirkner and Heinz Zimmermann).
In a different area, yet sharing my philosophy of learning from data, I am working on machine learning techniques for text mining and am interested in understanding data collected on the Web, in order to model the behavior of individuals.

Working Papers

• The abstracts of all of my Stern working papers including links to their postscript and pdf version are in www-psych.stanford.edu/~andreas/Research/Research.html

Publications

• The book on Decision Technologies for Financial Engineering (this page containts the abstracts of all the papers) is now available from World Scientific and from most bookstores, e.g., www.barnesandnoble.com (for USD 22.40).
• The full vita (resume), in case you want to find a specific citation (also as Word97 , pdf , and ps ).
• My current bibliographyfile for BibTex/LaTex contains a number of entries that might save you typing them in.

Conference CF99

• I organized the conference Computational Finance 1999 (CF99).It took place at Stern on January 6-8, 1999. It is sponsored by the New York University Salomon Center, the Center for Research on Information Systems, and the Department of Statistics and Operation Research. Blake LeBaron (University of Wisconsin at Madison and Brandeis) and Andrew Lo (MIT Sloan School of Management) was program co-chairs, and Yaser Abu-Mostafa (California Institute of Technology) was general chair.
• With more than three hundred attendees, CF99 successfully brought together decision-makers and strategists from financial firms and academics from finance, information systems, computer science, statistics, economics and other disciplines. The conference emphasized in-depth analysis of new techniques and their evaluation in comparison to established approaches. CF99 began with one full day of tutorials, presenting recent advances in several areas of computational finance to the diverse group of attendees. The two days of the main conference consisted of invited talks and keynotes, as well as the presentation of fully refereed papers. About ten percent of the submitted manuscripts were acceptance as talks. A poster session in the evening turned out to be an exciting forum for the exchange of new ideas. The conference proceedings will be published by MIT Press during summer 1999. The next conference, Computational Finance 2000, will take place at the London Business School.

• Computer-Based Systems for Management Support
  This course is for undergraduates and should be the first IS class students take at Stern. Concepts about the Internet and the World Wide Web, modern telecommunication, information technology, and management information systems are presented and critically discussed. The concepts encountered in this course range from the nature of model building and order-of-magnitude calculations, to electronic commerce and social issues of new technologies.
• C20.0001.05, 3 credits.
• Time: Tue Thu 9:50 - 11:30 UC65
The grades (both the points of the final exam and the overall lettergrade) were posted December 22nd.

Spring 1999

• Data Mining in Finance: Computer Intensive Methods for Financial Modeling and Intelligent Data Analysis. (The 14-page description is also available as pdf, as ps, or as doc.)
  This is a new course that develops links between Information Systems, Statistics and Finance. It covers the foundations of modern modeling, knowledge discovery and data mining techniques, as well as specific methods including neural networks, hidden Markov models, and model-based clustering. This course highlights the assumptions of different model classes, stresses the critical evaluation and comparison with established methods, and offers techniques for interpreting the results. Applications include current problems in finance such as understanding and managing risk, and building and evaluating trading models. Computer assignments are based on MATLAB. Possibilities of in-depth group projects in conjunction with major financial firms exist. Towards the end of the semester, several Wall Street practitioners present their perspectives on data mining in finance. Professor Weigend was awarded a 1998 NYU Curricular Development Challenge Grant to develop this course. Data Mining in Finance is taught in the Spring, for the first time this semester.
• B20.3355 (IS) = B90.3355 (SOR), 3 credits.
• Audience: MS/Statistics-Mathematical Finance, MS/IS, MBA/Financial Engineering Track, any PhD.
• Time: Mon Wed  7:00pm - 8:20pm (changed from earlier time). Room: K-MEC 2-70 (changed from earlier room).
• Statistical Artificial Intelligence and Decision Support Systems
  This course presents key papers from statistical machine learning, data mining and knowledge discovery. Applications to learning from data in business and finance are discussed. Active participation by the students is expected.
• B20.3389, 3 credits.
• Time: TBA.
• Prerequisite: PhD program or consent of instructor.

• Applying Hidden Markov Experts to Finance: Predicting Densities and Discovering Hidden States (June 10 - 12, Yale University).
  Talk at The Tenth Yale Workshop on Adaptive and Learning Systems.
• Beyond Data Mining: Predictive Modeling and Decision Making (June 19, Yorktown Heights, NY).
  Colloquium at IBM Research (Host: Scott Kirkpatrick).
• Modeling Hidden Structure in Financial Markets (June 22, Windows on the World, One World Trade Center, NYC).
  Talk at MATLAB Financial Engineering Conference.
• Discovering and Evaluating Hidden Structure in Time Series (June 23, Setauket, NY).
  Colloquium at Renaissance Technologies.
• Blending Supervised With Unsupervised Learning in Time Series Analysis and Prediction (June 29 - July 2, Technion, Haifa, Israel).
  Talk at Nonlinear Time Series for Prediction and Control .
• Data Mining in Finance (July 6, Zurich, Switzerland).
  One-day executive course at Risk Management Team / Schweizerisches Institut fuer Banken und Finanzen.
• Learning From Data: Physicists on Wall Street (July 10, Karlsruhe, Germany).
  Talk at Economics Department (Wirtschaftswissenschaften; Host: Gholamreza Nakhaeizadeh).
• Neural Networks and Time Series: The First Decade. (July 13, Cambridge, MA).
  Physics of Computation Seminar (MIT Media Lab; Host: Neil Gershenfeld).
• Learning From Data and Evaluating Predictive Models (August 27, Mitre Tech, Tysons Corner, VA).
  Talk at Financial Flows Workshop.
• Knowledge Discovery Through Predictive Models (August 31, Marriot Marquis, NYC).
  Invited talk at KDD Workshop on Data Mining in Finance.
  Also, the paper Discovering Technical Traders in the T-Bond Futures Market will be presented at the main KDD (Knowledge Discovery and Data Mining) Conference (August 27-31, NYC).
• Predicting Stock Return Densities Comparing GARCH, State Space, Gated Experts and Hidden Markov Experts (September 1, Rio de Janeiro, Brasil).
  Talk at Catholic University PUC-Rio (Host: Carlos Pedreira).
• Learning From Data: Building, Evaluating, and Understanding Models (September 2-4, Rio do Janeiro, Brasil).
  Keynote talk at ICDM (International Conference on Data Mining) .
• Discovering and Evaluating Hidden Structure in Financial Markets (September 16, Windows on the World, One World Trade Center, NYC).
  Workshop/tutorial at the annual conference of IAFE (International Association of Financial Engineers).
  The paper Analyzing Trading Behavior in the Treasury Bond Futures Marketwill be presented at the main conference (September 17-18).
• Learning and Combining Multiple Models for Density Forecasts (October 2, Berlin, Germany).
  Talk at Humboldt Universitaet (Statistik und Oekonometrie, Wirtschaftswissenschaften).
• Knowledge Discovery in Finance (October 6, Berlin, Germany).
  Half-day course at the Autumn School on Distributed Information Systems.

Info

Research

CF99 Computational Finance

Teaching

1991 Santa Fe Competition

• If you would like to know a little more about the kind of research I do, in quite general terms here are two interviews I participated in. They are written by Clive Davidson, and appeared in Olsen & Associates Views from the Frontier: The New Science, and The New Tools.
• And if you would like to know a little more about what I thought about 10 years ago, here is the Stanford Convocation speech I gave to the incoming graduate students. I'm actually surprised how much I still believe in it!

Any feedback is welcome: aweigend@stern.nyu.edu. Thank you.