Event(s) on May 2008

• 14/5/2008

  題目:	construction of large-scale linkage disequilibrium maps based on constrained unidimensional nonnegative scaling model
  講員:	Mr. LIAO Haiyong, Mathematics, Hong Kong Baptist University, HKSAR
  時間/地點:	10:30  -  11:30 FSC1217, Fong Shu Chuen Building
  摘要:	In my presentation, I will talk about the construction of large-scale linkage disequilibrium (LD) maps based on unidimensional nonnegative scaling model. The proposed constrained scaling model is firstly transformed to an unconstrained model,then solved by steepest descend method. The algorithm is implemented in PC Clusters at Hong Kong Baptist University. The LD maps are constructed for four populations from Hapmap data sets with chromosomes of several ten thousand single nucleotide polymorphisms (SNPs). The similarities and dissimilarities of the LD maps are studied and analyzed. Computational results are also reported to show the effectiveness of the method using parallel computation.

  
  

• 16/5/2008

  題目:	Forecasting Value-At-Risk with a ParsimoniousPortfolio Spillover GARCH (PS-GARCH) Model
  講員:	Prof. Michael McAleer, Economics, University of Western Australia, Australia
  時間/地點:	11:00  -  12:00 RRS905, Conference Room
  摘要:	Accurate modelling of volatility (or risk) is important in finance, particularly as it relates to the modelling and forecasting of Value-at-Risk (VaR) thresholds. As financial applications typically deal with a portfolio of assets and risk, there are several multivariate GARCH models which specify the risk of one asset as depending on its own past as well as the past behaviour of other assets. Multivariate effects, whereby the risk of a given asset depends on the previous risk of any other asset, are termed spillover effects. In this paper we analyse the importance of considering spillover effects when forecasting financial volatility. The forecasting performance of the VARMA-GARCH model of Ling and McAleer (2003), which includes spillover effects from all assets, the CCC model of Bollerslev (1990), which includes no spillovers, and a new Portfolio Spillover GARCH (PS-GARCH) model, which accommodates aggregate spillovers parsimoniously and hence avoids the so-called curse of dimensionality, are compared using a VaR example for a portfolio containing four international stock market indices. The empirical results suggest that spillover effects are statistically significant. However, the VaR threshold forecasts are generally found to be insensitive to the inclusion of spillover effects in any of the multivariate models considered.

  
  

• 22/5/2008

  題目:	The Solution of the Distance Geometry Problem for Protein Modeling
  講員:	Prof. Zhijun Wu, Department of Mathematics, Iowa State University, USA
  時間/地點:	11:30  -  12:30 RRS905, Conference Room
  摘要:	A well-known problem in protein modeling is the determination of the structure of a protein with a given set of inter-atomic or inter-residue distances obtained from either physical experiments or theoretical estimates. A general form of the problem is known as the distance geometry problem in mathematics, the graph embedding problem in computer science, and the multidimensional scaling problem in statistics. The problem has applications in many other scientific and engineering fields as well such as sensor network localization, image recognition, and protein classification. We describe the formulations and complexities of the problem in its various forms, and introduce a geometric buildup approach to the problem. Central to this approach is the idea that the coordinates of the atoms in a protein can be determined one atom at a time, with the distances from the determined atoms to the undetermined ones. The determination of each atom requires the solution of a small system of distance equations, which can usually be obtained in constant time. Therefore, in ideal cases, the coordinates of n atoms can be determined by a geometric buildup algorithm with O(n) distances in O(n) computing time instead of O(n2) distances in O(n2) computing time as required by a conventional singular-value decomposition algorithm. We present the general algorithm and discuss the methods for controlling the propagation of the numerical errors in the buildup process, for determining rigid vs. unique structures, and for handling problems with inexact distances (distances with errors). We show the results from applying the algorithm to a set of model protein problems with varying degrees of availability and accuracy of the distances and justify the potential use of the algorithm in protein modeling practice.