This course develops non-parametric methods for dependent data with applications to Finance, including optimal portfolio selection in a general setting. The goals of the course include the revision of the theory on sequential investment strategies for financial markets in a multi-period discrete time framework; description of non-parametric methods for the prediction of time series; revision of basic principles of universal consistent estimation; efficiency analysis of predictions based on combination of estimates; and empirical illustration of these methods
Professor of Computer Science and Information Theory at Budapest University of Technology and Economics. László Györfi received his Phd in Mathematical Sciences from the Hungarian Academy of Sciences (1988). Professor Györfi’s areas of research include stochastic approximation, pattern classification, non-parametric density, regression and entropy estimation, prediction of time series, multiple access communication, source coding and empirical portfolio selection. He has over one hundred publications in the main statistical and mathematical journals and is co-author of eight monographs.
May 7, 9 and 11: 16:00-19:00
May 15, 16 and 18: 16:00-18:00
Universidad Carlos III de Madrid (Getafe Campus)
Calle Madrid, 126
Building 15, Room 15.0.15 (Map)
*Fee covers the course and accompanying materials. The registration fee for students currently enrolled in a Ph.D. program is reduced to 300 Euros. Students may apply for scholarships by sending the Application Form at firstname.lastname@example.org, indicating as subject “SSECO-scholarship”. Course attendees need to arrange and pay for their own lodging.
Scholarships Applications will not be considered after April 30, 2012.