Title: | Shape invariant modeling pricing kernels and risk aversion |
Speaker: | Prof. Wolfgang Hardle, Centre for Applied Statistics & Economics, School of Business and Economics, Humboldt-Universität zu Berlin, Germany |
Time/Place: | 11:30 - 12:30 FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University |
Abstract: | Pricing kernels play a major role in quantifying risk aversion and investors’ preferences. Several empirical studies reported that pricing kernels exhibit a common pattern across different markets. Mostly visual inspection and occasionally numerically summarize are used to make comparison. With increasing amount of information updated every day, the empirical pricing kernels can be viewed as an object evolving over time. We propose a systematic modeling approach to describing the evolution of the empirical pricing kernels. The approach is based on shape invariant models. It captures the common features contained in the shape of the functions and at the same time characterizes the variability between the pricing kernels based on a few interpretable parameters. The method is demonstrated with the European options and returns values of DAX index. |
Title: | Applications of Semismooth Newton method for nonsmooth variational problems |
Speaker: | Kazufumi Ito, Department of Mathematics, North Carolina State University, USA |
Time/Place: | 11:30 - 12:30 FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University |
Abstract: | Applications and analysis of the semismooth Newton method for non-smooth equations associated with a general class of nonsmooth variational problems will be presented. The semismmooth Newton method is a generalized Newton method for a class of the Lipschitz but not $C^1$ equations in Banach spaces. It has been used to solve non-smooth equations that arise in the constrained optimization, variational inequality, contact and friction problems and mathematical finance, and numerical partial differential equations. In the case of the variational problem the nonsmooth equation is reduced from the necessary optimality condition for minimizing the cost functional involving $L^1$, $L^infty$ and $TV$ (total variation) norms and/or subject to the point-wise constraint of the solution and its gradient. In the image analysis and the control problem the $L^1$ norm is used to obtain the spike and impulsive solutions and $TV$ norm is used to capture the edge and the discontinuity in the image. A robust algorithm based on the primal and dual variable and the semismooth Newton is developed and analyzed. |
Title: | CMIV Colloquium: Prediction of protein-RNA binding sites by a random forest method with combined features |
Speaker: | Prof. Luonan Chen, Osaka Sangyo University, Japan |
Time/Place: | 17:00 - 18:00 DLB514, David C. Lam Building, Shaw Campus, Hong Kong Baptist University |
Abstract: | In this work, we propose a novel classification method to identify the RNA binding sites in proteins by combining a new interacting feature (interaction propensity) with other sequence and structure based features. Specifically, the interaction propensity represents a binding specificity of a protein residue to the interacting RNA nucleotide by considering its two-side neighborhood in a protein residue triplet. The sequence as well as the structure based features of the residues are combined together to discriminate the interaction propensity of amino acids with RNA. We predict RNA interacting residues in proteins by implementing a well-built random forest classifier. The experiments show that our method is able to detect the annotated protein-RNA interaction sites in a high accuracy. Our method achieves an accuracy of 84.5%, F-measure of 0.85, and AUC of 0.92 prediction of the RNA-binding residues for a dataset containing 205 non-homologous RNA-binding proteins, and also outperforms the existing methods in the comparison study. Furthermore, we provide some biological insights into the roles of sequences and structures in protein-RNA interactions by both evaluating the importance of features for their contributions in predictive accuracy and analyzing the binding patterns of interacting residues. |
Title: | Progress of the IIM for CFD problems with free boundary/moving interface |
Speaker: | Prof. Li Zhilin, Center for Research in Scientific Computation & Mathematics, North Carolina State University, USA |
Time/Place: | 11:30 - 12:30 FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University |
Abstract: | The Immersed Interface Method (IIM) is an efficient numerical method for solving interface, free boundary/moving interface problems, and problems on irregular domains. The IIM is motivated by the Peskin's Immersed Boundary (IB) method with better accuracy and more flexible to deal with different situations. The IIM is a sharp interface method that enforces jump conditions either exactly or approximately. In this talk, I will summarize some recent advances of the IIM, particularly, the applications to incompressible Stokes and Navier-Stokes equations with singular sources, discontinuous viscosity, irregular domains, and free boundary and moving interfaces using the augmented IIM. Applications include flow past cylinders, moving contact line problems, deformable moving interfaces, and incompressible interfaces in incompressible flows. |
Title: | DLS: High Accuracy Analysis of Finite Element Methods with applications for convection-diffusion problems |
Speaker: | Prof. Lin Qun, The Academy of Mathematics and Systems Sciences, Chinese Academy of Sciences, China |
Time/Place: | 16:30 - 17:30 (Preceded by Reception at 4:00pm) RRS905, Sir Run Run Shaw Building, HSH Campus, Hong Kong Baptist University |
Abstract: | Conventional accuracy analysis of finite element methods depends on the Schwarz inequality which yields the optimal accuracy for general meshes. However, for constructed meshes higher order of accuracy may be obtained without using the Schwarz inequality. Instead, we should use the “identity method” to recover the full accuracy. Especially for convection-diffusion problems involving small parameters, the traditional estimate certainly depends on the parameters and thus we can not obtain the optimal result. By using the “identity method” together with the properties of the target problems, we can obtain the optimal accuracy independent of the parameters. Moreover, the “identity method” only uses the integration by parts, thus most math postgraduate students may follow the approaches used. |
Title: | Two Data Mining Methods for SNP specific extraction and analysis |
Speaker: | Ms. Liu Yang, Department of Mathematics, Hong Kong Baptist University, Hong Kong |
Time/Place: | 10:00 - 11:30 FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University |
Abstract: | In this paper, two data mining methods were proposed to identify and analyze a set of relevant SNPs for the association between a disease and multiple marker genotypes. The first one is a subspace categorical clustering algorithm, which computes a weight for each SNP in the group of case samples and control samples, and uses the weights to identify the subsets of relevant SNPs that categorize these two groups. The second one is a classification shrunken centroid method, which can succinctly characterize each class (case and control) by shrinking each centroid with respect to the overall centroid by a certain threshold in a categorical manner, and select a set of SNPs associated with a particular disease. Both methods are applied in a genome-wide Parkinson disease data set. Results show that the proposed methods are suitable and useful tools to select relevant SNPs with genetic variations in a genome-wide association disease study. |
Title: | CMIV Colloquium: Gene-Gene interaction analysis for the genome-wide association studies: analysis of Korea Genome-wide Association data |
Speaker: | Prof. Taesung Park, Department of Statistics, Seoul National University, Korea |
Time/Place: | 11:00 - 12:00 DLB514, David C. Lam Building, Shaw Campus, Hong Kong Baptist University |
Abstract: | Gene-gene interaction is well recognized as playing an important role in understanding complex traits. Multifactor dimensionality reduction (MDR) was proposed by Ritchie, et al. (2001) to identify the multiple loci that affect disease susceptibility simultaneously. Although the MDR method has been widely used to detect gene-gene interactions in disease association studies, it cannot be easily applied to handle the whole genome association due to two limitations. First, the MDR procedure would require heavy computing in the genome-wide association research because the interaction effects are evaluated for all possible SNP combinations. Second, the evaluation is based on a cross-validation (CV) approach, and the single best SNP combination for each-order interaction is usually reported without declaring statistical significance. We propose a simple two-step strategy to unravel the gene-gene interactions in the framework of MDR approach for the whole genome association studies. In the first step, the marginal effects are tested for all SNPs, and the significant effects are reported. In the second step, we perform the second-order interactions via MDR analysis only for the SNP combinations of those SNPs which have significant marginal effects. Our proposed method is applied to Korea Association Resource Project (KARE) for 8800 subjects with 550K SNPs. |
Title: | Stepwise Searching in High-Dimensional Regression |
Speaker: | Prof. YAO Qiwei, Department of Statistics, The London School of Economics and Political Science, UK |
Time/Place: | 11:30 - 12:30 FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University |
Abstract: |
We investigate the classical stepwise forward and backward search
methods for selecting sparse models in the context of linear
regression with the number of candidate variables p greater than
the number of observations n. Two types of new information criteria
BICP and BICC are proposed to serve as the stopping rules in
the stepwise searches, since the traditional information criteria
such as BIC and AIC are designed for the cases with p |
Title: | DLS: Recent Progresses on Schrodinger Flows |
Speaker: | Prof. Weiyue Ding, Chinese Academy of Sciences, China |
Time/Place: | 11:30 - 12:30 (Preceded by Reception at 11:00am) RRS905, Sir Run Run Shaw Building, HSH Campus, Hong Kong Baptist University |
Abstract: | About 12 years ago the speaker and his colleague Wang, Youde, published a paper in which they proposed the notion of Schrodinger Flows inspired by the well-known Heisenberg model. This is an infinite dimensional Hamiltonnian system for maps from Riemannian manifolds into Kahler manifolds. In the past 10 years many research works have been done for Schrodinger Flows. In this lecture we will review some main results in the study of Schrodinger Flows, including local and global existence of the Cauchy problems; possible blow-up phenomenon; existence of self-similar solutions, time-periodic solutions and other special solutions. |
Title: | Numerical Simulation of Subsurface Multiphase Flow using Locally Conservative Finite Element Methods |
Speaker: | Prof. SUN Shuyu, Computational Transport Phenomena Laboratory Physical Sciences and Engineering Division, King Abdullah University of Science and Technology, Kingdom of Saudi Ara |
Time/Place: | 14:30 - 15:30 FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University |
Abstract: | Subsurface multiphase flow has important applications in petroleum reservoir engineering and environmental science. Modeling equation system of subsurface multiphase flow can be generally split into 1) an elliptic partial differential equation (PDE) for the pressure and 2) one or multiple convection dominated convection-diffusion PDE for the saturation or for the chemical composition. Accurate simulation of the phenomena not only requires local mass conservation to be retained in discretization, but it also demands steep gradients to be preserved with minimal oscillation and numerical diffusion. The heterogeneous permeability of the media often comes with spatially varied capillary pressure functions, both of which impose additional difficulties to numerical algorithms. To address these issues, we solve the saturation equation (or species transport equation) by discontinuous Galerkin (DG) method, a specialized finite element method that utilizes discontinuous spaces to approximate solutions. Among other advantages, DG possesses local mass conservation, small numerical diffusion, and little oscillation. The pressure equation is solved by either a mixed finite element (MFE) scheme or a Galerkin finite element method with local conservative postprocessing. In this talk, we will present the theory and numerical examples of this combined finite element approach for simulating subsurface multiphase flow. |
We organize conferences and workshops every year. Hope we can see you in future.
Learn MoreProf. M. Cheng, Dr. Y. S. Hon, Dr. K. F. Lam, Prof. L. Ling, Dr. T. Tong and Prof. L. Zhu have been awarded research grants by Hong Kong Research Grant Council (RGC) — congratulations!
Learn MoreFollow HKBU Math