Researchers (faculty and students) from the Centre for Computational Mathematics in Industry and Commerce regularly engage in collaborative research with others from a wide spectrum of fields (some examples are given below).
It could be the beginning of a significant inter-disciplinary research program. Below are some examples.
Hedge Funds Feel Pain
Recently, the New York Times reported that many hedge funds were having difficulty due to being long GM bonds, and short GM stock. Normally, this hedge works since, if bond prices go down, usually share prices go down, so that gain in the short position in GM stock counteracts the decrease in bond prices. However, due to a takeover bid, GM bonds were downgraded to junk, but GM share prices went up! Peter Forsyth is collaborating with the School of Accountancy's Ken Vetzal, graduate student Simon Clift, and financial software vendor Ito33 to develop more sophisticated computational finance models of the interaction between credit risk and share prices.
Probability of a Genetic Mutation for Colorectal Cancer (HNPCC) Given Family History
Wayne Oldford and MMath student Qing Li worked with medical researchers Dr. Roger Green and Dr. Ban Younghusband of Memorial University's Faculty of Medicine to develop a computational mathematical model combining Mendelian genetics and statistical survival data to produce the probability of carrying a mutation of colorectal cancer genes (hMSH2/hMLH1) based on the patient's family history. A basic program implementing the model is available. See Qing Li's master's research essay (PDF) for details.
Functional Magnetic Resonance Imaging (fMRI) Scans
Dr. Jeff Orchard has collaborated with Dr. Bruce Bjornson, a pediatric neurologist at the BC Childrens' Hospital. Bruce and his team do functional magnetic resonance imaging (fMRI) scans of people to view what part of their brain is activated by certain stimuli. Professor Orchard created an automated computer system that performs all the necessary processing (and there is lots of it) so that the results from the scan can be viewed within minutes after the scan. Previously, one had to wait overnight for the results.
Modelling Functional Genomics
In collaboration with Rafal Kustra (Public Health Sciences, University of Toronto), Romy Shioda and Mu Zhu have developed a factor analysis model for functional genomics. In this project, about 4000 genes from yeast are mapped to 369 functional categories using gene microarray expression data and gene ontology data. Rather than using traditional classification approaches, they model the problem using factor analysis and find the genes' classification using a variety of optimization approaches, including principal component analysis, semi-definite relaxation, and mixed-integer nonlinear optimization.
Decision Support for Health Informatics
Beginning in 2003, Romy Shioda and Mu Zhu have collaborated with researchers Jose Arocha, John Hirdes, and Trevor Smith of Applied Health Studies and with Kumaraswamy Ponnambalam of System Design Engineering to develop a decision support tool for health informatics. Professors Shioda and Zhu have been testing various data mining tools on the mental health minimum data set that is collected by mental health institutions throughout Ontario. This interdisciplinary project has also involved industry collaboration with Pattern Discovery Systems, a data analysis software company started by professors from University of Waterloo.