Home Page of Reg Kulperger

applied stochastic modeling and statistical inference for stochastic process
- flouresence and image correlation spectroscopy.
- use of a massively parallel computer in statistics
- stochastic modeling in forestry
empirical processes and randomly weighted empirical processes; statistical applications to tests of randomness and residuals; residuals for nonlinear time series
financial modeling, options, relation between discrete time and continuous time modeling; nonlinear time series; computational methods; pricing in incomplete markets and discrete time models
asymptotics
process bootstrapping; Gaussian processes, point processes and Markov processes

Conference Notes

Graduate students, PhD:

1992: John Braun: statistical inference for an interacting particle system model of a tumor growth pattern.
W. J. Braun, R. Kulperger and T. Lookman. (1990) Identification of an interacting two type particle system. In ``Stochastic Modelling in Biology,'' edited by P. Tautu, p401--411. World Scientific Publishing Co., New York.
Braun and Kulperger (1993) differential equtions for moments of an interacting particle process on a lattice. J. Math Biology, 31, 199--214.
1996: Alf Benn: a filtered Poisson process model of flouresence correlation spectroscopy. Alf is currently at the Toronto Dominion Bank, Toronto.
Benn and Kulperger (1996) Environmentrics, 7, 167--83, and Canadian Journal of Statistics (1997).
1998: Janusz Kawczak : a randomly weighted empirical process for residuals from autoregression; choice of a metric for weak convergence with unbounded random weights; empirical processes for ARCH residuals. Janusz is currently at North Carolina, Charlotte.
1999 : Pantelis Andreous: a randomly weighted empirical process for residuals from a linear regression model. weak convergence, numerical study of the process for tests of randomness of residuals. Pantelis is currently a statistician in Biostatistics and Epidemiology at Dalhousie University.
2000 : Gena Watteel: financial models, discrete and continuous time; nonparametric estimation of the Levy mixing measure. She is currently working for a private consulting company in Ottawa, Canada.
2007 : Alex Badescu, discrete time financial modeling. pricing in incomplete markets using GARCH models with non-normal driving noise, or using mixture GARCH models. Alex is Assistant Professor of Mathematics, University of Calgary.
2007 : Doug Woolford, supervised jointly with Professor Dave Stanford. Stochastic models of forestry and fire spread. After a postdoctoral position divided between University of Toronto, Forestry, and Simon Fraser University, Statistics, he has taken the position, starting July 2009, of Assistant Professor, Mathematics, at Wilfrid Laurier University.
2004 to 2008 : Taehan Bae is studying models of corporate default and ratings transitions. During 2008 to 2010 he is holding an NSERC Industrial Post Doctoral Fellowship at Algorithmics.

Current students:

I have some MSc students working on discrete time finance modeling and simulation questions. These include risk neutral option derivative pricing for ARCH type; Markov models of the returns process; estimation for nonlinear time series; empirical processes of residuals; penalized estimation.

There are MSc projects in modeling of forest fire starts and stochastic models of fire growth.

DOGS:

More information on my past and current dogs and related canine topics can be found by clicking on DOGS

The following two (Frodo and Dexter) did not work on modern statistical applications, but assisted in other ways.

Frodo and Dexter (August 1999).

froddex1.jpg (21512 bytes)

Dexter (1995-2001) died prematurely with a very aggressive form of cancer. Frodo (1993 - 2003) died in February 2003 at age 9 and 1/2. July 1, Frodo's birthday is a national holiday in Canada. It goes by two names, Canada Day and Frodo Day, although the second is not as well known nationally.

New Students:

medical imaging: UWO has a well known medical research community, including a group at the forefront of medical imaging headed by Aaron Fenster. I am looking for a student to work on some ideas of data reduction, classification and comparison of ultrasound and other imaging data. The initial method will be an application of spectral methods. We will make use of Fenster's computing facilities. As the work progresses, Fenster's group will also run experiments for us to supply us with the best possible data for the methods. His group develops the hardware, and are very interested in the participation of statisticians.

image correlation spectroscopy (ICS): this is a problem coming out the biochemistry work of Professor Nils Petersen of the department of Chemistry, UWO. I have some ideas of how we can exploit some higher moments and cumulants in a filtered Poisson model. These will, if they work in practice, yield directly some of the information that is currently only available in an indirect way. It has great potential in the study of proteins. We can also consider other ways in which one can exploit high tech computing facility in statistical problems.
I have various other questions about point processes, continuing work in empirical processes and financial modeling and data analysis questions that are of interest to graduate students.
Modeling questions in stochastic finance : (i) using discrete time returns models to evaluate risk neutral pricing of options; (ii) nonparametric methods for (i); stability questions related to (i); (iii) estimation of diffusion parameters in 1 dimensional SDE models using discrete time data; (iv) goodness of fit for SDE models in 1 dimensional SDE models; (v) estimation for nonlinear time series and residuals for nonlinear time series.

Other interests:

reading science fiction, Tolkien and mystery novels.
golf, hiking with my dogs, dog training methods
Star Trek

Poetry In Flanders Field (in Word Format)

In Flanders Fields in HTML

TestSecurehttp\SecTestTrial.htm

Stat3858\Exams\OldExams.htm

TBae (this is password protected)

04 Jan 2010