DEPT OF STATISTICS & ACTUARIAL SCIENCE
| Researcher : Chan WS |
| Project Title: | Living To 100 and Beyond Society of Actuaries International Symposium The Lee-Carter Model for Forecasting Mortality Revisited |
| Investigator(s): | Chan WS |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | URC/CRCG - Conference Grants for Teaching Staff |
| Start Date: | 01/2005 |
| Abstract: |
| N/A |
| Researcher : Chen A |
| Project Title: | Stochastic modelling of interacting branching systems |
| Investigator(s): | Chen A |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | General Research Fund (GRF) |
| Start Date: | 08/2006 |
| Completion Date: | 07/2009 |
| Abstract: |
| (1) The main objects of this proposal are as follows: a To validate an appropriate methodology that provides new techniques for investigating interacting branching systems. In particular, (i). the powerful probability approaches, such as the random change technique,will be developed in order to analyze properties of interacting branching systems; (2) (ii). the useful (one-dimensional) Resolvent Decomposition Theorem, refined in Chen and Renshaw (1990, 1993), will be generalized to the multi- dimensional case and facilitate the analysis of jointly interacting branching systems; (iii).the analytic approach, particularly methods in differential equations and special functions, will be further developed in order to analyze the partial differential equations arising from the study, especially for weakly and pair-wise interacting branching systems. (3)b. To investigate the basic properties of interacting branching systems. We expect to find solutions to the following open problems in the course of this project, which will greatly increase our understanding of interacting branching systems. (iv). We intend to obtain the extinction probability, the mean extinction time and the conditional mean extinction time of strongly interacting branching systems. (4)(v). We intend to obtain the explosion probability and the mean explosion time for strongly interacting branching systems. (vi). We intend to investigate and resolve the effect of immigration and emigration on the extinction and explosion properties for weakly, pair-wise, and strongly interacting branching systems. (vii).The limiting-conditional and quasi-stationary distributions for all types of interacting branching systems are unexplored at present. Both will be studied in detail. We expect to make considerable progress in this respect. (5)c. To establish and extend research collaboration, both nationally and internationally, in the study of interacting branching systems. Analyzing interacting branching systems is a challenging task and thus research collaboration is essential in order to achieve the aims of this ambitious project.(6) The results of the study will have a major long-term impact on mathematics and science, and in particular to biologists. It will promote interest, in particular, across the academic community within HK and Mainland China. Given that all the investigators of this proposal are leaders in their research fields and have strong track records, the aims of the study, while certainly ambitious, are readily achievable as long as funding and other strong support can be provided. We are confident that the project can be brought to fruition within the specified time period. This research will improve our understanding of interacting branching systems and spur further research worldwide in this important area. This in turn will enhance the reputation of HK and Mainland China in the branching systems research and cognate areas. |
| Researcher : Chen P |
| List of Research Outputs |
| Chen P. and Yang H., Pension Funding Problem with Regime Switching Geometric Brownian Motion Assets and Liabilities, Applied Stochastic Models in Business and Industry. Wiley, 2010, 26 (2): 125-141. |
| Researcher : Cheung KC |
| Project Title: | Portfolio choice under the cumulated prospect theory |
| Investigator(s): | Cheung KC |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | Seed Funding Programme for Basic Research |
| Start Date: | 06/2009 |
| Abstract: |
| The main thrust of my research proposal is to study optimal portfolio problems in which asset returns are dependent on each other, under Kahneman and Tversky’s (cumulative) prospect theory, featuring S-shaped utility functions and probability distortions. The main goals include (1) obtain conditions such that the optimal rank ordering of the allocations can be induced; (2) study how the optimal rank ordering of the allocations depends on the dependency structure. In the classical single-period utility-maximization portfolio problem, asset returns are usually modeled as independent random variables for analytic convenience. Some recent works have been done to relax this unrealistic assumption. As pointed out by Cheung and Yang (2004), ignoring the actual dependency structure by simply assuming independence may lead to a wrong rank ordering of the optimal allocations. Therefore, it is important to study how the dependency structure affects the optimal portfolio mix. Group hierarchical structure is one of the analytically tractable dependence structures. Intuitively, the returns of those securities belonging to the same group tend to move together, exhibiting a certain degree of positive dependence. However, assets belonging to different groups will have less dependent price movements. One possible way to model this phenomenon is to use the latent variable approach. Such approach was employed in Bauerle (1997) and Bauerle and Muller (1998) in modeling insurance risks. Let Xi denote the return of asset i, which belongs to group j. We could model Xi as a function of three independent latent variables: Gj, Yi, and Z. Here Z represents a global factor modeling the impact of the entire financial market. This factor will affect the returns of all assets. The random variable Gj is the group specific factor that affects only the asset returns in group j, but it has no effect on assets in other groups. The factor Yi is the individual factor that only affects Xi. Thus, we may express Xi as f(Gj, Yi, Z). In this case, all the asset returns are dependent on each other, through the common global market factor Z. Moreover, asset returns belonging to the same group will have stronger dependence since they are affected by a common group specific factor. As closed-form solutions are not available in general, the objective is to obtain sufficient conditions for ordering the optimal allocations within each group and across different groups. The solution method will involve the use of stochastic orders. The optimization criterion is to not the traditional expected von Neumann Morgenstern utility maximization. Instead, I propose putting the portfolio selection problem under the framework of Kahneman and Tversky’s (cumulative) prospect theory, featuring S-shaped utility (value) functions and probability distortions. In 1970s, Kahneman and Tversky (1979) proposed the prospect theory (PT) for decision making under uncertainty, incorporating human emotions and psychology into their theory. Later, Tversky and Kahneman (1992) modified slightly the PT to the cumulated prospect theory (CPT) so as to make the theory consistent with the first-order stochastic dominance. According to Jin and Zhou (2008), the key elements of Kahneman and Tversky’s Nobel–prize-winning theory are 1. a reference point in wealth that defines gains and losses; 2. a value function (which replaces the notion of utility function), concave for gains and convex for losses, and steeper for losses than for gains; 3. a probability distortion that is a nonlinear transformation of the probability scale, which enlarges a small probability and diminishes a large probability. There have been ongoing efforts in incorporating the PT or CPT into portfolio choice, for example Benartzi and Thaler (1995), Shefrin and Statman (2000), Levy and Levy (2004), Bassett,Koenker, Kordas (2004), Gomes (2005), DeGiorgi and Post (2005), and Jin and Zhou (2008). Nonetheless, none of them studied the effect of the dependence structure of asset returns on the portfolio choice. Therefore, it is the aim of the present research proposal to fill this gap. |
| Project Title: | Conditional comonotonicity and its application in actuarial science and financial economics |
| Investigator(s): | Cheung KC |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | General Research Fund (GRF) |
| Start Date: | 10/2009 |
| Abstract: |
| 1) Develop the theory of conditional comonotonicity systematically: (1) basic properties: for instance, whether conditional comonotonicity is preserved under different operations (e.g. mixing, convolution, etc); (2) characterizations of conditional comonotonicity (Cheung (2007b) is in this direction, Cheung (2008d) characterized comonotonicity using convex ordering) and its relationship with various stochastic orderings (e.g. the convex ordering, the supermodular ordering); (3) study the role played by conditional comonotonicity in the study of dependence structures and dependence measures. As demonstrated in Cheung (2007b), conditional comonotonicity is equivalent to local comonotonicity, it may be possible to introduce a new stochastic ordering comparing the "degree of comonotonicity" through comparing the underlying the sigma-fields being conditioned on. 2) Study the application of conditional comonotonicity in optimal portfolio problems. 3) Study the application of conditional comonotonicity in optimal insurance contract problems/best approximation problems. 4) Study the application of conditional comonotonicity in conditional risk measures. 5) Study other possible applications in actuarial science, financial economics, probability theory, etc. |
| Project Title: | 14th International Congress on Insurance: Mathematics and Economics (IME2010) Determining Dependence Structure from the Distribution of the Sum |
| Investigator(s): | Cheung KC |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | URC/CRCG - Conference Grants for Teaching Staff |
| Start Date: | 06/2010 |
| Completion Date: | 06/2010 |
| Abstract: |
| N/A |
| List of Research Outputs |
| Cheung K.C., Comonotonic convex upper bound and majorization, Insurance: Mathematics and Economics. 2010, 47: 154-158. |
| Cheung K.C., Fortis Chair Best Paper Award IME-2009, The K.U.Leuven - Fortis Chair. 2010. |
| Cheung K.C., Optimal reinsurance revisited - a geometric approach, ASTIN Bulletin. 2010, 40: 221-239. |
| Dong J., Cheung K.C. and Yang H., Upper comonotonicity and convex upper bounds for sums of random variables, Insurance: Mathematics and Economics. 2010, 47: 159-166. |
| Researcher : Chung YK |
| List of Research Outputs |
| Chung Y.K., Hu Y. and Fung T.W.K., Evaluation of DNA mixtures from database search, Biometrics. The International Biometric Society, 2010, 66: 233-238. |
| Researcher : Dong Y |
| List of Research Outputs |
| Dong Y. and Lee S.M.S., A new notion of data depth based on goodness-of-fit tests, Joint Statistical Meetings, Washington, DC. 2009. |
| Researcher : Fu J |
| List of Research Outputs |
| Fu J. and Yang H., Robust Replication of Volatility Derivatives for Time-changed Lévy Processes, 14th International Congress on Insurance: Mathematics and Economics. 2010. |
| Researcher : Fung TWK |
| Project Title: | 56th Session of the International Statistical Institute (Lisboa 2007) Estimating variance and covariance parameters by generalized estimating equations for credibility models |
| Investigator(s): | Fung TWK |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | URC/CRCG - Conference Grants for Teaching Staff |
| Start Date: | 08/2007 |
| Abstract: |
| N/A |
| Project Title: | Transmission disequilibrium test under imprinting for quantitative traits based on case-parents trios and for qualitative traits when only one parent is available |
| Investigator(s): | Fung TWK |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | General Research Fund (GRF) |
| Start Date: | 10/2007 |
| Abstract: |
| 1) To construct a valid test (or tests) for genomic imprinting in quantitative traits locus (QTL), based on case-parent trios under the QTDT framework; to extend QTDT incorporating imprinting; and/or to construct the test for qualitative traits when only one parent is available. 2) To choose the right constant c for the QTDT related test statistics for quantitative traits constructed in (1) such that the tests attain a high statistical power. 3) To determine whether all these tests still work when Hardy Weinberg equilibrium (HWE) is violated. |
| Project Title: | Statistical genomics: DNA mixtures and genomic imprinting |
| Investigator(s): | Fung TWK |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | Croucher Senior Research Fellowships in Natural Sciences, Technology and Medicine |
| Start Date: | 09/2008 |
| Abstract: |
| Refer to hard copy |
| Project Title: | Joint Statistical Meetings 2009 Analysis of Longitudinal Data Using Partial Linear Models with Quadratic Inference Functions |
| Investigator(s): | Fung TWK |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | URC/CRCG - Conference Grants for Teaching Staff |
| Start Date: | 08/2009 |
| Completion Date: | 08/2009 |
| Abstract: |
| N/A |
| Project Title: | Research Output Prize |
| Investigator(s): | Fung TWK, Hu Y |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | Research Output Prize (in Faculty) |
| Start Date: | 12/2009 |
| Abstract: |
| To identify and recognize the best research outputs in different faculties. |
| List of Research Outputs |
| Chung Y.K., Hu Y. and Fung T.W.K., Evaluation of DNA mixtures from database search, Biometrics. The International Biometric Society, 2010, 66: 233-238. |
| Fung T.W.K. and Xu X., Estimation and robustness of linear mixed models in credibility context , Reseach Report. Department of Statistics and Actuarial Science, HKU, 2010, 469: 1-27. |
| Yang C.T., Fung T.W.K. and Tam T.W.M., Alcohol study on blood concentration estimation: Reliability and applicability of Widmark formula on Chinese male population , Legal Medicine. Elsevier Ireland Ltd., 2009, 11: 163-167. |
| Zang Y., Fung T.W.K. and Zheng G., Asymptotic powers for matched trend tests and robust matched trend tests in case-control genetic association studies, Computational Statistics and Data Analysis. Elsevier B.V., 2010, 54: 65-77. |
| Zang Y., Fung T.W.K. and Zheng G., Simple algorithms to calculate asymptotic null distributions of robust tests in case-control genetic association studies in R, Journal of Statistical Software. The American Statistical Association, 2010, 33(8): 1-24. |
| Zhou J., Ding J., Fung T.W.K. and Lin S., Detection of parent-of-origin effects using general pedigree data, Genetic Epidemiology. Wiley-Liss, Inc, 2010, 34: 151-158. |
| Researcher : Hu Y |
| List of Research Outputs |
| Chung Y.K., Hu Y. and Fung T.W.K., Evaluation of DNA mixtures from database search, Biometrics. The International Biometric Society, 2010, 66: 233-238. |
| Researcher : Kwan CW |
| List of Research Outputs |
| Chen J.Y., Tse E.Y.Y., Lam T.P., Li D.K.T., Chao D.V.K. and Kwan C.W., Doctors' personal health care choices: a cross sectional survey in a mixed public/private setting., LKS Faculty of Medicine School of Public Health Inauguration. 2009. |
| Lou V.W., Chi I., Kwan C.W. and Leung A.Y.M., Changes in the functioning health of nursing home residents with moderate to severe cognitive impairement, 2010 Joint World Conference on Social Work & Social Development: The Agenda. June 10-14, Hong Kong, 2010, 164. |
| Lou V.W., Chui E.W.T., Leung A.Y.M., Kwan C.W., Chi I. and Leung Wong E.K.S., Psychological factors affecting long-term care utilization in Hong Kong, A symposium on Community-based Long-term Care in Hong Kong: From Theory to Practice. Hong Kong, 2009. |
| Lou V.W., Chui E.W.T., Leung A.Y.M., Kwan C.W., Chi I. and Leung E., The role of psychosocial factors in long-term care utilization among elderly Chinese, The Gerontological Society of America 62nd Annual Scientific Meeting. Atlanta, GA, The Gerontological Society of America, 2009, 111. |
| Lou V.W., Kwan C.W., Chi I. and Leung A.Y.M., Willingness to provide care of family caregivers for community-dwelling older adults: Trends in Hong Kong 2004, 2006, 2008, 2010 Joint World Conference on Social Work & Social Development: The Agenda. June 10-14, Hong Kong, 2010, 163. |
| Researcher : Kwan KM |
| List of Research Outputs |
| Kwan K.M. and Yang H., Dependent Insurance Risk Model: Deterministic Threshold, Communications in Statistics: Theory and Methods. Taylor & Francis, 2010, 39 (5): 765-776. |
| Researcher : Lam KF |
| List of Research Outputs |
| Cheung Y.B. and Lam K.F., Three estimates of the association between linear growth failure and cognitive ability, Tropical Medicine and International Health. Blackwell Publishing Ltd., 2009, 14(9): 1-5. |
| Lam K.F., Xu Y. and Cheung T.L., A multiple imputation approach for clustered interval-censored survival data, Statistics in Medicine. John Wiley & Sons, Ltd, 2010, 29: 680-693. |
| Lam K.F., Editor, Hong Kong Medical Journal. Hong Kong Academy of Medicine, 2010. |
| Lam K.F., Statistical Advisor for The Journal of Orthopaedic Surgery, In: Prof. David Fang, Hong Kong, Hong Kong Academy of Medicine Press, 2010. |
| Lam T.P., Wong J.G.W.S., Ip M.S.M., Lam K.F. and Pang S.L., Psychological wellbeing of interns in Hong Kong: What causes them stress and what helps them, Medical Teacher . Taylor & Francis, 2010, 32: e120-e126. |
| Li K.Y., Wong M.C.M., Lam K.F. and Schwarz E., Marginal approach in analyzing sequential cross-sectional oral health survey data, Journal of Dental Research. 2009, 88 (Spec Iss B): 311 (PAPF/APR). |
| Mok T.M.Y., Huang F.P., Ip W.K., Lo Y., Wong F.Y., Chan E.Y.T., Lam K.F. and Xu D., Serum levels of IL-33 and soluble ST2 and their association with disease activity in systemic lupus erythematosus, Rheumatology. Oxford, 2010, 49: 520-527. |
| Ng F.H., Wong S.Y., Lam K.F., Chu W.M., Chan P., Ling Y.H., Kng C.P.L., Yuen W.C., Lau Y.K., Kwan A. and Wong B.C.Y., Famotidine is inferior to Pantoprazole in preventing recurrence of aspirin-related peptic ulcers or erosions, Gastroenterology. Elsevier, 2010, 138: 82-88. |
| Siu D.C.W., Lau C.P., Lee S.W.L., Lam K.F. and Tse H.F., Intravenous diltiazem is superior to intravenous amiodarone or digoxin for achieving ventricular rate control in patients with acute uncomplicated atrial fibrillation, Critical Care Medicine. MD Consult, LLC, 2009, 37(7): 2174-2179. |
| Wang S., Yiu K.H., Mok T.M.Y., Ooi C.G.C., Khong P.L., Mak H.K.F., Lau C.P., Lam K.F., Lau W.C.S. and Tse H.F., Prevalence and extent of calcification over aorta, coronary and carotid arteries in patients with rheumatoid arthritis, Journal of Internal Medicine . 2009, 266: 445-452. |
| Yiu K.H., Wang S.L., Mok T.M.Y., Ooi C.G.C., Khong P.L., Mak H.K.F., Lam K.F., Lau W.C.S. and Tse H.F., Pattern on atherosclerosis for coronary, carotid and aortic arteries calcification in rheumatoid arthritis: a multidetector CT study, Journal of American College of Cardiology. 2009, 53: A432. |
| Yiu K.H., Wang S., Mok T.M.Y., Ooi C.G.C., Khong P.L., Lau C.P., Lai W.W., Wong L.Y., Lam K.F., Lau W.C.S. and Tse H.F., Role of circulating endothelial progenitor cells in patients with rheumatoid arthritis with coronary calcification, Journal of Rheumatology . 2010, 37: 529-535. |
| Researcher : Lam Y |
| List of Research Outputs |
| Liang X., Lam Y. and Li Z., Optimal replacement policy for a general geometric process model with d-shock , Research Report. Department of Statistics and Actuarial Science, HKU, 2010, 470: 1-30. |
| Researcher : Lee H |
| List of Research Outputs |
| Lam K., Yu P.L.H. and Lee H., A margin scheme that advises on when to change required margin, European Journal of Operational Research. Elsevier B.V., 2010, 207: 524-530. |
| Lee H. and Yu P.L.H., Distance-based tree models for ranking data, Computational Statistics and Data Analysis. Elsevier, 2010, 54: 1672-1682. |
| Researcher : Lee H |
| List of Research Outputs |
| Lam K., Yu P.L.H. and Lee H., A margin scheme that advises on when to change required margin, European Journal of Operational Research. Elsevier B.V., 2010, 207: 524-530. |
| Lee H. and Yu P.L.H., Distance-based tree models for ranking data, Computational Statistics and Data Analysis. Elsevier, 2010, 54: 1672-1682. |
| Researcher : Lee SMS |
| Project Title: | An omnibus procedure for robust depth-based confidence regions |
| Investigator(s): | Lee SMS |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | General Research Fund (GRF) |
| Start Date: | 01/2007 |
| Completion Date: | 12/2009 |
| Abstract: |
| (1) Problems: Conventionally confidence regions for parameter vectors are built upon sampling distributions of approximate pivotal quantities. For a particular real-life application there often exist more than one choices of pivotal quantities and methods for estimating their distributions. The claimed quality of each such confidence procedure derives somewhat from our subjective belief in the validity of its governing conditions, which are difficult to ascertain in reality. Although adaptive and robust methods have emerged in the literature to optimize the choice of confidence procedure and guard against mis-specification of governing conditions respectively, their scope is confined to parametric contexts or special forms of pivotal quantities, and the parameter often restricted within one-dimensional spaces. In general situations where different confidence procedures are available of both the parametric and non-parametric types, based on pivotal quantities of different functional forms or of different dimensions, an omnibus approach that successfully integrates these diverse possibilities is desirable in practice. Ad hoc approaches based on, for example, formation of simple unions or convex hulls of individual confidence regions provide a plausible solution but often yield regions of awkward shapes, having over-conservative coverages and bestriding irrelevant candidate parameter values. The problem remains unsolved in a formal and thorough way. (2) Proposed solution: This project seeks a solution in the form of an omnibus confidence procedure, which defines a confidence region by depth-based calibration with reference to an amalgam of simulated generalized pivotal quantities obtained from different confidence procedures. It systematically integrates the separate confidence regions that would have been obtained by individual confidence procedures. The region constructed can be carefully tuned to optimize coverage accuracy or maintain a conservative confidence level. Both theoretical and empirical properties of the proposed procedure are investigated under sufficiently general regularity conditions and for a variety of practical examples, respectively. (3) Significance and value: The project introduces a novel and practically viable confidence procedure to the standard statistical toolkit. It integrates and improves upon existing standard methods when, as is common in reality, external information falls short of identifying a single best choice among the standard methods. The resulting region enjoys desirable accuracy and robustness properties. Being entirely data-driven, its content adapts faithfully to the real situation and is particularly informative about the location of the target parameter. |
| Project Title: | An adaptive combination of parametric and nonparametric bagged predictors |
| Investigator(s): | Lee SMS |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | General Research Fund (GRF) |
| Start Date: | 01/2008 |
| Abstract: |
| Compare theoretically the parametric and nonparametric bagging procedures under a correctly specified parametric model. Examine theoretically the effects of model mis-specification on the performance of parametric bagging. Develop a hybrid bagging procedure which combines the parametric and nonparametric versions of bagging in an optimal way. Investigate the theoretical and empirical properties of the hybrid procedure. |
| Project Title: | A new class of data depths based on goodness-of-fit tests |
| Investigator(s): | Lee SMS |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | General Research Fund (GRF) |
| Start Date: | 01/2009 |
| Abstract: |
| 1) Formulate a systematic construction of data depth, endowed with a “representativeness” interpretation, in both finite- and infinite-dimensional settings. 2) Compare the new class of depths with previous definitions devised for multivariate data. Particular attention is given to the governing properties that have conventionally been demanded of multivariate data depth. 3) Explore new features peculiar to the proposed class of data depths, especially those in connection with their renewed interpretation as a measure of “representativeness”. 4) Investigate the empirical properties of the new depths in different contexts of statistical applications, with emphasis on data of the infinite-dimensional type. |
| Project Title: | 1st Institute of Mathematical Statistics - Asia Pacific Rim Meeting Improving Coverage Accuracy of Block Bootstrap Confidence Intervals |
| Investigator(s): | Lee SMS |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | URC/CRCG - Conference Grants for Teaching Staff |
| Start Date: | 06/2009 |
| Completion Date: | 07/2009 |
| Abstract: |
| N/A |
| List of Research Outputs |
| Cheung K.Y. and Lee S.M.S., Bootstrap variance estimation for Nadaraya quantile estimator, Test. 2010, 19: 131 – 145. |
| Dong Y. and Lee S.M.S., A new notion of data depth based on goodness-of-fit tests, Joint Statistical Meetings, Washington, DC. 2009. |
| Dong Y. and Lee S.M.S., Depth functions as measures of representativeness, Research Report. Department of Statistics and Actuarial Science, HKU, 2010, 474: 1-33. |
| Lee S.M.S., Hybrid confidence regions based on data depth, Research Report. Department of Statistics and Actuarial Science, HKU, 2010, 473: 1-42. |
| Researcher : Li G |
| List of Research Outputs |
| Li G. and Li W.K., Least Absolute Deviation Estimation For Unit Root Processes With Garch Errors, In: Peter C. B. Phillips, Econometric Theory. Cambridge, 2009, 25: 1208–1227. |
| Li M., Li G. and Li W.K., Score tests for long memory in stochastic volatility, Research Report. Department of Statistics and Actuarial Science, HKU, 2010, 467: 1-28. |
| Researcher : Li M |
| List of Research Outputs |
| Li M., Li G. and Li W.K., Score tests for long memory in stochastic volatility, Research Report. Department of Statistics and Actuarial Science, HKU, 2010, 467: 1-28. |
| Researcher : Li WK |
| Project Title: | Statistical inference for time series with threshold moving average structure |
| Investigator(s): | Li WK, Tong H |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | General Research Fund (GRF) |
| Start Date: | 01/2007 |
| Abstract: |
| (1) Long term impact: Threshold models, in particular the threshold autoregressive models have played an important role in the development of nonlinear time series analysis since its introduction by Tong (1978) and Tong & Lim (1980). Unlike the linear situation, research in threshold time series concentrates mainly on the autoregressive type models. The corresponding threshold moving average (TMA) process has been given little attention in the literature through the years, apart from some mainly theoretical developments. Given the importance of the moving average models in linear time series analysis it can be expected that a threshold moving average model should have a similarly important role in nonlinear time series analysis. Research in time series with a threshold moving average component is long over due. Such time series processes should provide better understanding of various nonlinear phenomenona that are happening in the real world. Towards this goal the investigators propose to undertake the following tasks of investigation on time series models with a threshold moving average structure. (2) Objective 1: Statistical inference for the pure TMA model. This will include estimation methodologies and large sample properties of the estimators such as consistency and asymptotic distributions. Testing and identification procedures will be developed for testing and identifying the presence of threshold structure, possibly under presence of conditional heteroscedasticity. Goodness of fit tests will be developed to check for model adequacy. Procedures for model selection such as methods of selecting model order and choosing between different models will be considered. (3) Objective 2: Develop new time series models with a threshold moving average (TMA) component; investigates their probabilistic structure and derive appropriate statistical inference methodologies. Some immediate generalizations of the basic threshold moving average model would include the threshold autoregressive and moving average (TARMA) model, vector threshold moving average (VTMA) model, and threshold moving average models with conditional heteroscedasticity. (4) Objective 3: Computer programmes will be developed to model and forecast time series with a TMA structure. Applications to the real world will be included using environmental, economic and financial data. Clearly, the objectives only reflect part of the possible developments with threshold moving averages. Many other developments, such as Bayesian analysis and semiparametric models are possible if time permits. |
| Project Title: | Statistical Inference for the Hyberbolic Generalized Conditional Autoregressive Heteroscedastic (HYGARCH) Model and Extensions |
| Investigator(s): | Li WK |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | General Research Fund (GRF) |
| Start Date: | 01/2009 |
| Abstract: |
| 1) Estimation methodology We shall start with the investigation of the asymptotic properties of the parameter estimates of the HYGARCH model. This focuses primarily on the quasi-maximum likelihood estimates (QMLE) but other types of estimates such as the least absolute deviation estimates may also be considered. The results here should be informative in obtaining similar results for extensions of HYGARCH models such as the threshold or the asymmetric power variants of the basic HYGARCH model. Simulations will be used to gauge the accuracy of the large sample results. It is anticipated that the QMLE may not be most effective under certain conditions and alternative estimation methods such as the Bayesian approach or the indirect estimator of Gourieroux et al. (1993) will be investigated. 2) Model diagnostics checks and hypotheses testing Model diagnostic checking is an important step in time series modelling. The project will derive goodness-of-fit tests for the HYGARCH models under various modelling approaches. This will result in useful extensions of classical portmanteau tests for ARCH models (Li and Mak, 1994) but may include other possibilities. As the HYGARCH model nests other models such as the FIGARCH and the IGARCH models statistical tests will be derived to compare these models. Because of nusiance parameters at the boundary of the parameter space, these tests may very likely possess a nonstandard large sample distribution which will be tabulated as far as possible. Simulations will be performed to gauge the empirical sizes and powers of the tests derived. 3) Extensions of the HYGARCH models New types of time series for volatility modeling will be developed based on the HYGARCH model. In particular, a threshold HYGARCH model will be defined and estimation methodology for this model will be developed. The new model should be capable in capturing the asymmetric behaviour of volatility during different market conditions. Asymptotic properties of the parameter estimates will be established as in Objective 1. Tests for the presence of threshold structure will be derived and their asymptotic distribution tabulated. The capability of the new model will be tested by applying it to real data. If time permits other extensions such as the asymmetric power HYGARCH will also be studied. |
| List of Research Outputs |
| Ching W.K., Siu K.T.K., Li L., Jiang H., Li T. and Li W.K., An Improved Parsimonious Multivariate Markov Chain Model for Credit Risk, Journal of Credit Risk . 2009, 5: 1-25. |
| Ching W.K., Siu K.T.K., Li L., Li T. and Li W.K., Modeling Default Data via an Interactive Hidden Markov Model, Computational Economics. 2009, 34: 1-19. |
| Jayawardena A... .W..., Xu P... and Li W.K., Rainfall Data Simulation by Hidden Markov Model and Discrete Wavelet Transfomation, Stochastic Environment Research and Risk Assessment. USA, Springer, 2009, 23: 863-877. |
| Kwan C.K., Li W.K. and Ng K.W., A Multivariate Threshold GARCH Model with Time Varying Correlations, Econometric Reviews. USA, Taylor & Francis, 2010, 29: 20-38. |
| Li G. and Li W.K., Least Absolute Deviation Estimation For Unit Root Processes With Garch Errors, In: Peter C. B. Phillips, Econometric Theory. Cambridge, 2009, 25: 1208–1227. |
| Li M., Li G. and Li W.K., Score tests for long memory in stochastic volatility, Research Report. Department of Statistics and Actuarial Science, HKU, 2010, 467: 1-28. |
| Lui G.C.S., Li W.K. and Lee J.H.W., Ensemble Kalman filter with nonlinear updating equation , Research Report. Department of Statistics and Actuarial Science, HKU, 2010, 468: 1-31. |
| Wat K.P., Yuen K.C. and Li W.K., Time series insurance risk models with dependence structures (presented by Kam Pui Wat, PhD student), International Conference on Actuarial and Financial Risks, Shanghai, China. 2010. |
| Yu P.L.H., Wu E.H.C. and Li W.K., Financial Data Mining Using Flexible ICA-GARCH Models, In: A B M Shawkat Ali and Yang Xiang, Dynamic and Advanced Data Mining for Progressing Technological Development: Innovations and Systemic Approaches. USA, IGI Global, 2010, 255-272. |
| Yuen K.C., Li J., Wat K.P. and Li W.K., On the compound binomial risk model with dividends and time-correlated claims (presented by Kam Pui Wat, PhD student), The 14th International Congress on Insurance: Mathematics and Economics, Toronto, Canada. 2010. |
| Researcher : Lui GCS |
| List of Research Outputs |
| Bao W.W., Lui G.C.S. and Leung K.M.Y., Acute toxicities of zince pyrithione alone and in combination with copper to marine autotrophic species, the 6th international Conference on Marine Pollution and Ecotoxicology, 31 May-3 June 2010, City University of Hong Kong, Hong Kong. 2010. |
| Bao W.W., Lui G.C.S. and Leung K.M.Y., Mixture toxicities of zinc pyrithione and copper to subtropical marine organisms: implications on deriving environmentally realistic water quality criteria, the 6th international Conference on Marine Pollution and Ecotoxicology, 31 May-3 June 2010, City University of Hong Kong, Hong Kong. 2010. |
| Bao W.W., Leung K.M.Y. and Lui G.C.S., Mixture toxicities of zinc pyrithione and copper to subtropical marine organisms: implications on deriving environmentally realistic water quality criteria, the SETAC Asia/Pacific 2010 Meeting, held during 4-7 June 2010 at Guangzhou, China. 2010. |
| Lui G.C.S., Li W.K. and Lee J.H.W., Ensemble Kalman filter with nonlinear updating equation , Research Report. Department of Statistics and Actuarial Science, HKU, 2010, 468: 1-31. |
| Researcher : Ng KW |
| Project Title: | Compositional Data Analysis Workshop (CoDaWork'03) Compositional Hypotheses of Subcompositional Stability and Specific Perturbation Change and Their Testing |
| Investigator(s): | Ng KW |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | URC/CRCG - Conference Grants for Teaching Staff |
| Start Date: | 10/2003 |
| Abstract: |
| N/A |
| Project Title: | Further Properties and New Applications for the family of Nested Dirichlet Distributions |
| Investigator(s): | Ng KW, Tian G |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | Small Project Funding |
| Start Date: | 08/2009 |
| Abstract: |
| The family of Dirichlet distributions has been the primary family that is defined on the limited subspace of an n-dimensional space where all n coordinates sum to unity, called the (closed) simplex of the space; note that if we only consider the first (n-1) coordinates as free variables, called the open simplex. It appears in many applications including modeling of compositional data (such as percentage of contents in chemistry and material composition in rocks, etc.), Bayesian analysis, statistical genetics, Pearson systems of curves, non-parametric inference, distribution-free tolerance intervals, multivariate analysis, order statistics, reliability theory, probability inequalities, probabilistic constrained programming models, limit laws, delivery problems, stochastic processes, and other areas. Applications in these areas required extensions of the family in different directions to suit different purposes. The earlier ones are the Liouville family and the generalized Liouville family. The former has been discussed extensively in Fang, Kotz and Ng, (1990, Chapter 6) with particular applications in modeling compositional data, nonparametric prediction of lifetimes, survival functions and multiple Type I error of analysis of variance. The latter was developed by Rayens and Srinivasan (1994) with application to compositional data analysis. In connection with Bayesian computation for contingency tables with incomplete cell-counts, the family of Nested Dirichlet Distributions was briefly introduced by Tian, Ng and Geng (2003) with its stochastic representation for simulation purpose. This widely broadened family has two sets of parameters, one being the same as in the original Dirichlet family and the additional set consisting of (n-1) non-negative parameters. When the parameters in the additional set are all zero, this sub-family reduces to the Dirichlet family. Upon the need of applications in large-sample likelihood inferences and small-sample Bayesian inferences for incomplete categorical data, more properties of the broadened family were obtained by Ng, Tang, Tian and Tan (2009), including the mixed (or raw) moments and mode. However, the important and vital components of a standard distribution theory for this new family are still missing, and various potential applications of the family to the analysis of incomplete categorical data need to be investigated further. The missing properties include the marginal and conditional distributions, cumulative distribution function (and its complement – the survival distribution), moment generating function (and its companion – the characteristic function), and the distribution of a linear combination of the variables for the multivariate distribution. The purpose of the project is four-fold: (i) To find the marginal and conditional distributions of this new family of distributions and other missing properties; (ii) to establish the exact null distribution for the sphericity test with the aid of the marginal distribution of the Nested Dirichlet Distributions (iii) to investigate new applications in fitting competing-risks model, analyzing incomplete categorical data and evaluating cancer diagnosis tests, and (iv) to do simulation studies for comparing with the existing methodology. It's well known that the problem of finding the marginal and conditional distribution for a new family is not straightforward and thus needs special techniques case by case. Also generating random variables from a conditional distribution depends on special properties that are to be developed. |
| List of Research Outputs |
| Konstantinides D.G., Ng K.W. and Tang Q., The probabilities of absolute ruin in the renewal risk model with constant force of interest, Journal of Applied Probability. Applied Probability Trust, 2010, 47: 323-334. |
| Kwan C.K., Li W.K. and Ng K.W., A Multivariate Threshold GARCH Model with Time Varying Correlations, Econometric Reviews. USA, Taylor & Francis, 2010, 29: 20-38. |
| Ng K.W., Inversion of Bayes formula for events, Research Report. Department of Statistics and Actuarial Science, HKU, 2010, 466: 1-9. |
| Ng K.W., Tang M.L., Tian G. and Tan M., The Nested Dirichlet Distribution and Incomplete Categorical Data Analysis, Statistica Sinica . 2009, 19(1): 251-271. |
| Tan M., Tian G. and Ng K.W., Bayesian Missing Data Problems: EM, Data Augmentation and Non-iterative Computation, Chapman & Hall/CRC (Monographs in Biostatistics Series), Boca Raton, USA. 2009. |
| Tian G., Tan M., Ng K.W. and Tang M.L., A Unified Method for Checking Compatibility and Uniqueness for Finite Discrete Conditional Distributions, Communication in Statistics: Theory and Methods. 2009, 38(1): 115-129. |
| Tian G., Tang M.L., Yuen K.C. and Ng K.W., Further properties and new applications for the nested Dirichlet distribution, Computational Statistics and Data Analysis. Science Direct, 2010, 54(2): 394-405. |
| Tian G., Ng K.W., Li K.C. and Tan M., Non-iterative Sampling-based Bayesian Methods for Identifying Changepoints in the Sequence of Cases of Haemolytic Uraemic Syndrome, Computational Statistics and Data Analysis. 2009, 53(9): 3314-3323. |
| Researcher : Siu KTK |
| List of Research Outputs |
| Ching W.K., Leung H.Y., Jiang H., Sun L. and Siu K.T.K., A Markovian Network Model for Default Risk Management, , International Journal of Intelligent Engineering Informatics . 2010, 1: 104-124. |
| Ching W.K., Siu K.T.K., Li L., Jiang H., Li T. and Li W.K., An Improved Parsimonious Multivariate Markov Chain Model for Credit Risk, Journal of Credit Risk . 2009, 5: 1-25. |
| Song N., Ching W.K., Siu K.T.K., Fung S.L. and Ng K.P., Option Valuation Under a Multivariate Markov Chain Model, In: L. Yu, Y. Song, W. Ching, S. Wang and K. Lai, Proceedings of CSO2010, Huangshan,. IEEE Computer Society Proceedings, 2010, 1: 177-181. |
| Yiu C., Liu J..., Siu K.T.K. and Ching W.K., Optimal Portfolios with Regime-Switching and Value-at-Risk Constraint, Automatica. 2010, 46: 979-989. |
| Researcher : Tian G |
| Project Title: | Accelerating the Quadratic Lower-Bound (QLB) Algorithm via Optimizing Shrinkage Parameter |
| Investigator(s): | Tian G |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | Seed Funding Programme for Basic Research |
| Start Date: | 04/2010 |
| Abstract: |
| Logistic regression model and Cox proportional hazard model are two important models in applied statistics and are widely used in biomedical studies. Because of the property of quadratical convergence, the Newton-Raphson algorithm or the Fisher scoring algorithm is usually employed to find maximum likelihood estimates (MLEs) of the parameters of interest. However, the two methods have drawbacks such as requiring tedious calculation of matrix inverse of the observed information matrix (i.e., negative Hessian) or the expected information matrix at each iteration. Furthermore, the Newton-Raphson algorithm does not necessarily increase the log-likelihood, leading even to divergence sometimes (Cox and Oakes, 1984, p.172). For example, Bohning and Lindsay (1988, p.645-646) provided a simple example of a concave objective function in which the Newton-Raphson method does not converge. In this project, we will further show that the Newton-Raphson and Fisher scoring algorithms do not work for logistic regression with the cancer remission data (Lee, 1974). If we could create a structure of missing data, EM-type algorithms (Dempster et al., 1977; Meng and Rubin, 1993) were the best choice as they possess the ascent property guaranteeing the monotone convergence at a linear rate. However, the EM-type algorithms may not apply to generalized linear models (e.g., logistic regression and log-linear models) and Cox proportional hazard models due to the absence of a missing-data structure. Therefore, for problems in which the missing-data structure does not exist or is not readily available, minorization-maximization (MM) algorithms (Lange et al., 2000; Hunter & Lange, 2004) are often useful alternatives. As a special case of MM algorithms, the quadratic lower-bound (QLB) algorithm (Bohning & Lindsay, 1988) is an optimization device which monotonically increases the log-likelihood at each iteration. However, like EM-type algorithms, the QLB algorithm may suffer from excruciatingly slow convergence, especially for complicated problems or high-dimensional data. Motivated by the 'working parameter' idea (Meng & van Dyk, 1997) in the acceleration of the EM algorithm, we will develop in this project a novel 'shrinkage parameter' approach to accelerate the QLB algorithm while maintaining its simplicity and stability (i.e., monotonic increase in log-likelihood). The objective of this project is five-fold: (i) To develop a 'shrinkage parameter' approach to accelerate the original QLB algorithm; (ii) to provide some theoretical justifications and to establish an optimal QLB algorithm; (iii) to generalize the optimal QLB algorithm to penalized problems and to investigate some properties of convergence in this new setting; (iv) to apply the optimal QLB algorithm to fit logistic regression model and Cox proportional hazard model; and (v) to illustrate the proposed methods by a cancer remission dataset and a primary biliary cirrhosis dataset. |
| List of Research Outputs |
| Fang H.B., Tian G., Li W. and Tan M., Design and Sample Size for Evaluating Combinations of Drugs of Linear and Log-linear Dose-response Curves, Journal of Biopharmaceutical Statistics. 2009, 19(4): 625-640. |
| Gao W., Shi N...Z., Tang M...L., Fu L...Y. and Tian G., Unified generalized iterative scaling and its applications, Computational Statistics and Data Analysis. 2010, 54: 1066-1078. |
| Ng K.W., Tang M.L., Tian G. and Tan M., The Nested Dirichlet Distribution and Incomplete Categorical Data Analysis, Statistica Sinica . 2009, 19(1): 251-271. |
| Tan M., Tian G. and Ng K.W., Bayesian Missing Data Problems: EM, Data Augmentation and Non-iterative Computation, Chapman & Hall/CRC (Monographs in Biostatistics Series), Boca Raton, USA. 2009. |
| Tan M., Tian G. and Tang M.L., Sample Surveys with Sensitive Questions: A Non-randomized Response Approach, The American Statistician. 2009, 63(1): 9-16. |
| Tang M.L., Tian G., Tang N.S. and Liu Z., A new non-randomized multi-category response model for surveys with a single sensitive question: Design and analysis, Journal of the Korean Statistical Society. Elsevier B.V., 2009, 38: 339-349. |
| Tian G., Tan M., Ng K.W. and Tang M.L., A Unified Method for Checking Compatibility and Uniqueness for Finite Discrete Conditional Distributions, Communication in Statistics: Theory and Methods. 2009, 38(1): 115-129. |
| Tian G., Tang M.L., Yuen K.C. and Ng K.W., Further properties and new applications for the nested Dirichlet distribution, Computational Statistics and Data Analysis. Science Direct, 2010, 54(2): 394-405. |
| Tian G., Ng K.W., Li K.C. and Tan M., Non-iterative Sampling-based Bayesian Methods for Identifying Changepoints in the Sequence of Cases of Haemolytic Uraemic Syndrome, Computational Statistics and Data Analysis. 2009, 53(9): 3314-3323. |
| Tian G., Fang H.B., Liu Z. and Tan M.T., Regularized (bridge) logistic regression for variable selection based on ROC criterion, Statistics and Its Interface. International Press, 2009, 2: 493-502. |
| Tian G., Fang H.B., Tan M., Qin H. and Tang M.L., Uniform Distributions in a Class of Convex Polyhedrons with Applications to Drug Combination Studies, Journal of Multivariate Analysis. 2009, 100(8): 1854-1865. |
| Researcher : Tong H |
| Project Title: | Distinguished Visiting Professors in the Department of Statistics and Actuarial Science |
| Investigator(s): | Tong H |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | Distinguished Research Achievement Award |
| Start Date: | 05/2005 |
| Abstract: |
| N/A |
| List of Research Outputs |
| Tong H., Threshold models in time series analysis-30 years on, Research Report. Department of Statistics and Actuarial Science, HKU, 2010, 471: 1-36. |
| Xia Y.C. and Tong H., Feature matching in time series modelling, Research Report. Department of Statistics and Actuarial Science, HKU, 2010, 472: 1-38. |
| Researcher : Wat KP |
| List of Research Outputs |
| Wat K.P., Yuen K.C. and Li W.K., Time series insurance risk models with dependence structures (presented by Kam Pui Wat, PhD student), International Conference on Actuarial and Financial Risks, Shanghai, China. 2010. |
| Yuen K.C., Li J., Wat K.P. and Li W.K., On the compound binomial risk model with dividends and time-correlated claims (presented by Kam Pui Wat, PhD student), The 14th International Congress on Insurance: Mathematics and Economics, Toronto, Canada. 2010. |
| Researcher : Xu X |
| List of Research Outputs |
| Fung T.W.K. and Xu X., Estimation and robustness of linear mixed models in credibility context , Reseach Report. Department of Statistics and Actuarial Science, HKU, 2010, 469: 1-27. |
| Researcher : Xu Y |
| List of Research Outputs |
| Lam K.F., Xu Y. and Cheung T.L., A multiple imputation approach for clustered interval-censored survival data, Statistics in Medicine. John Wiley & Sons, Ltd, 2010, 29: 680-693. |
| Researcher : Yang CT |
| List of Research Outputs |
| Yang C.T., Fung T.W.K. and Tam T.W.M., Alcohol study on blood concentration estimation: Reliability and applicability of Widmark formula on Chinese male population , Legal Medicine. Elsevier Ireland Ltd., 2009, 11: 163-167. |
| Researcher : Yang H |
| Project Title: | Absolute ruin probability in a jump diffusion model |
| Investigator(s): | Yang H |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | Small Project Funding |
| Start Date: | 11/2007 |
| Abstract: |
| Absolute ruin probability was first considered in Gerber (1971, 1979). He gave an explicit expression for the absolute ruin probability in the compound Poisson risk model with debit interest when claims are exponentially distributed. Later on, Dassios and Embrechts (1989) studied the absolute ruin probability in the compound Poisson risk model with debit interest when claims are exponentially distributed using the martingale approach. Absolute ruin probabilities in a piecewise-deterministic Markov process risk model with credit and debit interest have been discussed in Embrechts and Schmidli (1994). They derived explicit expressions for the absolute ruin probabilities with exponential claims using the Laplace transform method. Cai (2007) reconsidered absolute ruin problems in the compound Poisson risk model with debit interest by defining an expected discounted penalty function at absolute ruin. He gave the explicit expression for the expected discounted penalty function at absolute ruin with exponential claims by a differential equation approach. A review of absolute ruin probabilities can be found in Asmussen (2000). This project considers the compound Poisson insurance risk model perturbed by diffusion with investment. We assume that the insurance company can invest its surplus in both a risky asset and the risk-free asset according to a fixed proportion. If the surplus is negative, a constant debit interest rate is applied. The absolute ruin probability function satisfies a certain integro-differential equation. We will derive the integro-differential equation. In various special cases, closed form solutions can be expected, and numerical illustrations will be provided. An alternative way is to use the technique of stochastic control to obtain the integro-differential equation satisfied by the absolute ruin probability. Using this equation, we derive the asymptotical forms of the absolute ruin probability when the distributions of claim sizes are heavy-tailed and light-tailed. We then decompose the absolute ruin probability into the sum of two probabilities, where one is the probability that absolute ruin is caused by a claim and the other one is the probability that absolute ruin is caused by oscillation. We will derive the integro-differential equations and asymptotical results for the two probabilities as well. We will illustrate the results using some numerical examples. Asmussen, S. (2000). Ruin Probabilities. World Scientific, Singapore. Cai, J. (2007). On the time value of absolute ruin with debit interest. Advances in Applied Probability, 39(2), in press. Dassios, A. and Embrechts, P. (1989). Martingales and insurance risk. Stochastic Models, 5, 181-217. Embrechts, P. and Schmidli, H. (1994). Ruin estimation for a general insurance risk model. Advances in Applied Probability, 26, 404-422. Gerber, H.U. (1971). Der Einfluss von Zins auf die Ruinwahrscheinlichkeit. Bulletin of the Swiss Association of Actuaries, 71, 63-70. Gerber, H. U. (1979). An Introduction to Mathematical Risk Theory. S.S. Huebner Foundation Monograph Series No.8, R. Iwin, Homewood, IL. |
| Project Title: | Risk Management of Equity-Linked Insurance Products |
| Investigator(s): | Yang H |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | General Research Fund (GRF) |
| Start Date: | 01/2009 |
| Abstract: |
| 1) Develop appropriate methodologies and tools for measuring and understanding the risk inherent from some modern insurance products with embedded options, such as, equity-linked products, guaranteed annuity options, and variable annuities, etc. 2) Introduce theoretically sound and practically useful models for pricing, hedging and reserving these products. 3) Investigate the impact of model dynamics for three key risk factors underlying these products, namely, the share market risk, the interest rate risk and the mortality risk on the pricing, hedging and reserving of these products. 4) Investigate and develop appropriate risk measures for these products. 5) Explore the state of the art of the modern pricing and hedging technologies in financial mathematics to deal with the pricing hedging issues of these modern insurance products and highlight the interplay between financial mathematics and actuarial science. |
| Project Title: | Option Pricing and ALM in Regime Switching Models |
| Investigator(s): | Yang H, Yin GG |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | General Research Fund (GRF) |
| Start Date: | 01/2010 |
| Abstract: |
| 1) Develop a trinomial tree method to price European, American and exotic options under regime switching models. 2) Investigate some problems in the ALM context under regime switching models. |
| Project Title: | The 7th. Symposium on Risk Management and Cyber-Informatics (RMCI 2010) Pricing Options and Equity-Indexed Annuities in a Regime-switching Model by Trinomial Tree Method |
| Investigator(s): | Yang H |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | URC/CRCG - Conference Grants for Teaching Staff |
| Start Date: | 06/2010 |
| Completion Date: | 07/2010 |
| Abstract: |
| N/A |
| List of Research Outputs |
| Chen P. and Yang H., Pension Funding Problem with Regime Switching Geometric Brownian Motion Assets and Liabilities, Applied Stochastic Models in Business and Industry. Wiley, 2010, 26 (2): 125-141. |
| Dong J., Cheung K.C. and Yang H., Upper comonotonicity and convex upper bounds for sums of random variables, Insurance: Mathematics and Economics. 2010, 47: 159-166. |
| Elliott R., Siu T.K. and Yang H., Filtering a Markov Modulated Random Measure, IEEE Transactions on Automatic Control. USA, IEEE Control Systems Society, 2010, 55 (1): 74-88. |
| Fu J. and Yang H., Robust Replication of Volatility Derivatives for Time-changed Lévy Processes, 14th International Congress on Insurance: Mathematics and Economics. 2010. |
| Gerber H.U., Shiu E. and Yang H., An Elementary Approach to Discrete Models of Dividend Strategies, Insurance; Mathematics and Economics. Netherlands, ELSEVIER, 2010, 46 (1): 109-116. |
| Gerber H.U., Shiu E. and Yang H., Crossing Time of Annuities with Exponential Payment Rates, Bulletin of the Swiss Association of Actuaries. Switzerland, 2009, 2009 (2): 96-100. |
| Kwan K.M. and Yang H., Dependent Insurance Risk Model: Deterministic Threshold, Communications in Statistics: Theory and Methods. Taylor & Francis, 2010, 39 (5): 765-776. |
| Lin S., Tan K.S. and Yang H., Pricing Annuity Guarantees under a Regime-Switching Model, North American Actuarial Journal. USA, Society of Actuaries, 2009, 13 (4): 316-332. |
| Siu T.K. and Yang H., Nonparametric Bayesian Credibility, Australian Actuarial Journal. Australia, Institute of Actuaries of Australia, 2009, 15 (2): 209-230. |
| Siu T.K. and Yang H., Option Prices When the Regime-Switching Risk is Priced, Acta Mathematicae Applicatae Sinica. Springer, 2009, 25 (3): 369-388. |
| Wei J., Yang H. and Wang R., On the Markov-Modulated Insurance Risk Model with Tax, Blaetter DGVFM. Springer, 2010, 31 (1): 65-78. |
| Yang H., An elementary approach to discrete models of dividend strategies, Third International Gerber-Shiu Workshop . 2010. |
| Yang H., Pricing Options and Equity-Indexed Annuities in a Regime-switching Model by Trinomial Tree Method, 2010 International Conference on Insurance and Actuarial Science. 2010. |
| Yang H., Pricing Options and Equity-Indexed Annuities in a Regime-switching Model by Trinomial Tree Method, Workshop on Financial Mathematics & Applied Statistics. 2009. |
| Yuen F.L. and Yang H., Option Pricing in a Jump-diffusion Model with Regime-switching, ASTIN Bulletin. UK, Peeters, 2009, 39 (2): 515-539. |
| Yuen F.L. and Yang H., Option Pricing with Regime-switching by Trinomial Tree Method, Journal of Computational and Applied Mathematics. Netherland, Elsevier, 2010, 233 (8): 1821-1833. |
| Yuen F.L. and Yang H., Option Pricing with Tree Model in View of Hedging, International Conference on Actuarial and Financial Risks. 2010. |
| Yuen F.L. and Yang H., Option Valuation by a Self-Exciting Threshold Binomial Model, 14th International Congress on Insurance: Mathematics and Economics. 2010. |
| Yuen F.L. and Yang H., Pricing Asian Options and Equity-Indexed Annuities with Regime-switching by Trinomial Tree Method, North American Actuarial Journal. USA, Society of Actuaries, 2010, 14 (2): 17. |
| Researcher : Yin G |
| Project Title: | 2010 Eastern North American Region Spring Meeting Bayesian Phase I/II Drug-combination Trial Design in Oncology |
| Investigator(s): | Yin G |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | URC/CRCG - Conference Grants for Teaching Staff |
| Start Date: | 03/2010 |
| Completion Date: | 03/2010 |
| Abstract: |
| N/A |
| List of Research Outputs |
| Yin G., Associate Editor, Bayesian Analysis. 2010. |
| Yin G. and Nieto-Barajas L.E., Bayesian cure rate model accommodating multiplicative and additive covariates, Statistics and Its Interface. International Press, 2009, 2: 513-521. |
| Yin G., Bayesian phase I/II drug-combination trial design in oncology with adaptive randomization. Department of Biostatistics, The University of Texas, M. D. Anderson Cancer Center, Houston, TX, January, 2010., 2010. |
| Yin G., Bayesian phase I/II drug-combination trial design in oncology. ENAR Spring Meeting. New Orleans, LA, March, 2010., 2010. |
| Yin G. and Yuan Y., Rejoinder to the discussion of "Bayesian dose finding in oncology for drug combinations by copula regression", Journal of the Royal Statistical Society, Series C (Applied Statistics). Wiley-Blackwell, 2010, 59(3): 544-546. |
| Researcher : Yip PSF |
| Project Title: | 22nd World Congress of the International Association for Suicide Preventation (IASP) Social Burden of Suicide Risk in Hong Kong |
| Investigator(s): | Yip PSF |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | URC/CRCG - Conference Grants for Teaching Staff |
| Start Date: | 09/2003 |
| Abstract: |
| N/A |
| Project Title: | IBC 2004 XXIInd International Biometric Conference in Parallel with ASC 2004 Australian Statistical Conference A Unification Method of Estimating Population Size via Capture-recapture Experiments |
| Investigator(s): | Yip PSF |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | URC/CRCG - Conference Grants for Teaching Staff |
| Start Date: | 07/2004 |
| Abstract: |
| N/A |
| Project Title: | XVIII World Congress of the International Association for Suicide Prevention Suicide in Three Chinese Communities: China, Taiwan and Hong Kong |
| Investigator(s): | Yip PSF |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | URC/CRCG - Conference Grants for Teaching Staff |
| Start Date: | 09/2005 |
| Abstract: |
| N/A |
| Project Title: | HKU Overseas Fellowship Awards 2009-10 |
| Investigator(s): | Yip PSF |
| Department: | Soc Work & Social Administration |
| Source(s) of Funding: | HKU Overseas Fellowship Awards |
| Start Date: | 09/2009 |
| Abstract: |
| To visit the University of Melbourne, Australia to conduct collaborative projects on modeling of the contagious effect of mass media reporting and suicide, assessing the effect of exclusion period among life insured, monitoring and surveillance system for public health event, and geographical information system modeling and teaching innovation. |
| Project Title: | Outstanding Researcher Award 2008-2009 |
| Investigator(s): | Yip PSF |
| Department: | Soc Work & Social Administration |
| Source(s) of Funding: | Outstanding Researcher Award |
| Start Date: | 12/2009 |
| Abstract: |
| The Awards are intended to recognize, reward, and promote exceptional research accomplishments of academic and research staff. |
| Project Title: | A Comparative Study on Quality of Suicide Prevention Websites in Hong Kong, Taiwan, and Mainland China |
| Investigator(s): | Yip PSF |
| Department: | Soc Work & Social Administration |
| Source(s) of Funding: | Small Project Funding |
| Start Date: | 12/2009 |
| Abstract: |
| Increasing number of people is using the Internet to access health related, including suicide prevention related, information and services. Correspondingly, concern on quality of the web-based information is also growing. However, studies on quality of web-based information are imbalanced between information available in English and in Chinese, as well as between information about physical and mental health. In the west, there are continuing studies on evaluating quality of health websites and testing validity of criteria and indicators of content quality, through quantitative or qualitative or both research methods. These studies addressed a consistent conclusion that it’s important to educate health websites owners as well as consumers to be aware of the importance of content quality. Furthermore, some “action” initiatives have been developed. Guidelines or quality criteria for health related websites have been issued by organizations in America (Winker, et al., 2000) and European countries (eEurope, 2002). Quality label, as widely used HONcode, can be applied from the non-profit organization, the Health On the Net Foundation (HON). Special search engines or online catalog and index providing reliable and trustworthy results can be visited at MedHunt (Switzerland), Medical Matrix (United States), MedWebPlus (United States), HealthWeb (United States), and HealthInsite (Australia), etc. More recently, an automated quality assessment procedure (AQA)(Griffiths, Tang, Hawking, & Christensen, 2005) was designed by Australian researchers to automatically rank depression websites according to their evidence-based quality. On the other hand, studies on non-English websites, such as French (CISMeFteam, 2007; Darmoni, Thirion, J.P., Douyere, & Piot, 2001) and Japanese(Goto, et al., 2009; Nemoto, et al., 2007) websites, have also been published. However, studies on websites in the language of Chinese are rare. Among Mainland China, Hong Kong, and Taiwan, Taiwan shows the highest awareness of quality of online health related information, which also has the most websites obtaining HONcode among the three regions. One study in Taiwan(Sing-Ling & Sin-Kuo, 2005) developed an evaluation questionnaire for nursing websites and achieved a positive conclusion on its validation. One study in Hong Kong(Yan, 2008) explored Hong Kong Internet users’ online health information seeking behavior and their confidence on health information obtained. The researcher also summed up several major criteria for health websites from questionnaire survey. The study showed that 44% of 443 participants were uncertain about the reliability of the health information that they obtained from the Internet, whereas 41% of them thought the information was reliable. Compared with Hong Kong and Taiwan, quality of health related websites seems to be relatively poor. Two Chinese papers(Li, 2004; Su, Li, & Wan, 2006) reviewed websites of centers for disease control and prevention in different administrative levels in Mainland China. They found that most of these websites are disorganized, unable to determine source and last update of information, lacking information about authorship, lacking localized and original information, lacking interactivity, and not user-friendly. However, the evaluation of these two papers was based on authors’ subjective review rather than evidence-based criteria or indicators. Another study (Ou, Zhao, & Liao, 2006) introduced HONcode briefly to its readers in Mainland China and suggested Chinese health websites to adopt it. According to the study, although HONcode has already translated into Chinese, there are very few websites from Mainland China and Hong Kong but several websites from Taiwan obtaining the certification. If we divide health related websites into physical health related and mental health related, there are much fewer studies focusing on the latter in particular. In mental health field, quality of web-based information about depression is relatively often studied (Belcher & Holdcraft, 2001; Griffiths, et al., 2005; Nemoto, et al., 2007) and the quality was found to be generally poor. If we take a closer look at websites about suicide prevention, so far there is no study focusing on evaluation of quality of these websites at all. Several studies(Biddle, Donovan, Hawton, Kapur, & Gunnell, 2008; Recupero, Harms, & Noble, 2008) examined online suicide related information by searching through popular search engines and showed that some encouraging suicide websites can be easily searched out and have quite high rankings in the search result list. To reduce the potential harm to vulnerable people, filtering or censorship on pro-suicide information is practically inefficient and sensitive on ethical and legal issues. Therefore, researchers in suicide prevention field (Becker, Mayer, Nagenborg, El-Faddagh, & Schmidt, 2004; Bell, 2007; Mishara & Weisstub, 2007) widely suggested that it would be effective to create and promote more qualified suicide prevention information online to counterbalance potential harmful information that may encourage vulnerable people to commit suicide. Recently, we have conducted a comparative study on online suicide related information in simplified Chinese and English, which is under reviewing by the Journal of Clinical Psychiatry. We found that it is more notably lack of qualified and professional suicide prevention information in Chinese websites, compared with English websites. For further study, we suggested to examine quality of suicide prevention websites carefully and sum up valid criteria for the quality evaluation. In the proposed study, we will review websites about suicide prevention in particular and will include websites not only from Mainland China, but also from Hong Kong and Taiwan. Our first aim is to evaluate quality of web based information on suicide prevention in Chinese, including simplified Chinese and traditional Chinese, and conduct a comparison between the three regions. In addition, we aim to identify valid indicators of quality of suicide prevention websites in Chinese. |
| Researcher : Yu PLH |
| Project Title: | Development of Data Mining Labs using SAS Enterprise Miner Version 5 |
| Investigator(s): | Yu PLH |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | Run Run Shaw Research and Teaching Endowment Fund - Teaching Grants |
| Start Date: | 06/2007 |
| Abstract: |
| Teaching and learning data mining techniques at The University of Hong Kong: Data mining aims at revealing underlying structure and hidden patterns in large amounts of data. Lots of successful data mining applications can be found in business, medicine, science and engineering. Since the academic year 2002-2003, the Department of Statistics and Actuarial Science has been offering data mining courses in which a very popular commerical data mining software---SAS Enterprise Miner (EM) is used. SAS's clients using EM include HSBC, Hang Seng Bank, Hong Kong Jockey Club, CSL, etc. Through the partnership with SAS Limited, the Department can purchase EM at a very low academic rate and students can install a licensed EM at their home PC free of charge. To help students gain hands-on experience of using EM, more than 10 data mining lab demonstrations were designed and used in the computer laboratory sessions. As all the analyses in EM require creating a work-flow diagram using its point-and-click interface, each set of data mining labs consists of step-by-step guideline showing how to select the appropriate options, explanatory notes of each option and some selected screen snapshots. In the past few years, students found the demonstrations very useful in learning the data mining techniques. So far, the courses have been using EM Version 4 (EM4) and all the labs and lecture notes are designed based on it. As SAS EM Version 5 (EM5) has been launched for a few years and become popular in many major banks in Hong Kong and in the world, I plan to teach this version in the coming year. However, EM5 has a completely different user interface including data input process, function selection panels and output interfaces, etc. The whole software is written in Java instead of an SAS add-on module like EM4. Furthermore, EM5 provides more new data mining tools such as path analysis, model assessment methods, graphical exploration tools and enhanced features in various data mining techniques. This project aims at revising the existing data mining lab materials and writing additional labs for the new data mining techniques available only in EM5. To help students learn EM5 more efficiently, a quick reference guide in a web page format will also be prepared. |
| Project Title: | Variational Bayesian Approach to Factor Analysis |
| Investigator(s): | Yu PLH |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | Small Project Funding |
| Start Date: | 11/2008 |
| Completion Date: | 04/2010 |
| Abstract: |
| Factor analysis (FA) is a powerful multivariate analysis technique that uncovers the latent common characteristics (or factors) among a set of variables and has been widely used for dimension reduction, feature extraction, time series prediction and so forth. There has been a long-standing interest in Baysian treatment for FA since it can avoid the over-fitting problem of maximum likelihood (ML) method and lead to determinate the number of factors automatically. However, conventional Baysian treatment uses sampling-based techniques, which is computationally demanding and often limits its use to small-scale problems. In recent years, due to advances of computationally more efficient approximate inference methods such as Variational Bayesian (VB) that can even accommodate large-scale problems, Bayesian approaches achieves increasing applications. VB has been applied to FA model and mixtures of FA’s or related models in Beal (2003), Ghahramani and Beal (2000) and further investigated in Nielsen (2004). Compared with Bayesian information criterion (BIC), however Nielsen (2004, p. 58) observes that VB for FA (henceforth denoted as VBFA) tends to be under-fitting; Beal (2003, p.142–p.143) also observes that a mixture version of VBFA tends to penalize model complexity more heavily, i.e., choose a model with less number of components. As is well known, BIC is simply a crude approximation to Baysian evidence. Intuitively, it seems that VB should perform comparable or better than BIC. These two somewhat surprising observations motivates us to develop a better VB treatment for FA. We find two problems associated with VBFA: (1) the large sample limit of VBFA does not correspond to BIC and has heavier penalty than BIC. This means that VBFA disagrees with the general theoretical result on latent variable models developed in Attias (1999) that BIC emerges as a limit case of VB; (2) the performance of VBFA in low noise case is unsatisfactory as low noises deteriorate the performance of VBFA, resulting in unneeded factors cannot be effectively suppressed. Expectation-Maximization (EM) algorithm has been suggested for fitting FA, which is easy to implement and converges stably since its M-step is in closed form. Despite the advantage of simplicity and stability, convergence of EM is only linear and can be painfully slow. Recently, Zhao, Yu (PI) and Jiang (2008) propose a conditional maximization (CM) algorithm, which shares the same advantage of EM but possesses quadratic convergence. In addition to achieve faster convergence, CM provides more insight in a more sensible choice of prior in VB treatment. This project will investigate a new VBFA using alternative choice of prior for model parameters. We also consider the extension of new technique to mixture of FA's (Ghahramani and Beal, 2000 and Zhao and Yu (PI), 2008). |
| Project Title: | Two-Dimensional Probabilistic Principal Component Analysis |
| Investigator(s): | Yu PLH |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | Small Project Funding |
| Start Date: | 12/2009 |
| Abstract: |
| Dimension reduction is important in many disciplines such as botany, biology, bioinformatics, social sciences, economics, and engineering, because the interesting structure of the high dimensional data can generally display in a low dimensional space and thus more compact and meaningful representation of the data is required for visualization, exploration, modeling, etc. Probabilistic modeling for dimension reduction is a central research area in data mining, pattern recognition, machine learning and statistics. Compared with non-probabilistic counterparts, probabilistic models enables different sources of uncertainty inherent in the data to be well studied by means of probability theory. Thus statistical inference and Bayesian (or Variational Bayesian) methods could be performed and missing data could be handled in a systematic manner. Models could be extended to mixtures of such models to accommodate heterogeneous data, modified to accommodate discrete data and robustified to handle `outliers' by incorporating a heavy-tailed distribution such as a student t-distribution. Principal component analysis (PCA) (Jolliffe, 2002) is one of most popular techniques for dimension reduction. Due to the non-probabilistic nature of PCA, Moghaddam and Pentland (1997) formulated PCA in a probabilistic framework and Tipping and Bishop (1999) derived the probabilistic PCA (PPCA) from the classical linear latent variable model. PPCA is an important development of PCA since it inherits all the advantages as a probabilistic model and includes PCA as a special case. Notice that PPCA is simply formulated for 1D data (in which observations are in vector form). To apply PPCA for 2D data (in which observations are in matrix form), one ad hoc approach is to apply PPCA to the vectorized data. However, this might not obtain the result as expected because (1) the vectorization breaks the natural matrix structure, which may incur loss of the potentially more compact or useful representation (Ye et al., 2004), and (2) it is often not suitable for the data such as images or image patches, as the resulting 1D data (typically, over tens of thousands) by vectorization is easily trapped into the so-called curse of dimensionality and could be computationally prohibited even on modern computers (Xie et al., 2008). The main objective of this project is to develop a novel probabilistic model called two-dimensional PPCA (2DPPCA) to overcome the above problems when PPCA is applied to 2D data. |
| List of Research Outputs |
| Lam K., Yu P.L.H. and Lee H., A margin scheme that advises on when to change required margin, European Journal of Operational Research. Elsevier B.V., 2010, 207: 524-530. |
| Lee H. and Yu P.L.H., Distance-based tree models for ranking data, Computational Statistics and Data Analysis. Elsevier, 2010, 54: 1672-1682. |
| Yu P.L.H., Associate Editor, Computational Statistics & Data Analysis. 2010. |
| Yu P.L.H., Wu E.H.C. and Li W.K., Financial Data Mining Using Flexible ICA-GARCH Models, In: A B M Shawkat Ali and Yang Xiang, Dynamic and Advanced Data Mining for Progressing Technological Development: Innovations and Systemic Approaches. USA, IGI Global, 2010, 255-272. |
| Zhao J. and Yu P.L.H., A note on variational Bayesian factor analysis, Neural Networks. Elsevier Ltd., 2009, 22: 988-997. |
| Researcher : Yuen FL |
| List of Research Outputs |
| Yuen F.L. and Yang H., Option Pricing in a Jump-diffusion Model with Regime-switching, ASTIN Bulletin. UK, Peeters, 2009, 39 (2): 515-539. |
| Yuen F.L. and Yang H., Option Pricing with Regime-switching by Trinomial Tree Method, Journal of Computational and Applied Mathematics. Netherland, Elsevier, 2010, 233 (8): 1821-1833. |
| Yuen F.L. and Yang H., Option Pricing with Tree Model in View of Hedging, International Conference on Actuarial and Financial Risks. 2010. |
| Yuen F.L. and Yang H., Option Valuation by a Self-Exciting Threshold Binomial Model, 14th International Congress on Insurance: Mathematics and Economics. 2010. |
| Yuen F.L. and Yang H., Pricing Asian Options and Equity-Indexed Annuities with Regime-switching by Trinomial Tree Method, North American Actuarial Journal. USA, Society of Actuaries, 2010, 14 (2): 17. |
| Researcher : Yuen KC |
| Project Title: | Ruin analysis of some insurance risk model with heavy-tailed claims and risky investment |
| Investigator(s): | Yuen KC |
| Department: | Statistics & Actuarial Science |
| Source(s) of Funding: | Small Project Funding |
| Start Date: | 12/2009 |
| Abstract: |
| The study of heavy-tailed claims is one of the key topics in actuarial science and applied probability. It certainly helps in handling the adverse effects of possible catastrophic events. On the other hand, ss investment is an increasingly important element in insurance business, the study of insurance risk models with risky investment has drawn a great deal of attention in the past decade. In view of their practical importance, we shall consider some ruin problems for the renewal risk model with both heavy-tailed claims and risky investments. Ruin probability is the probability that the surplus process ever falls below zero. It is one of the main criteria in insurance and risk management. For example, this probability can be used to provide an early warning system for the guidance of an insurance project. The time at ruin is another random variable of interest in modern risk theory. For the renewal risk model with heavy-tailed claims and risky investments, the project objective aims at developing asymptotic results for the finite-time probability, the infinite-time ruin probabilities, and the Laplace transform of the time of ruin. As an application, we shall make use of the asymptotic results to determine the optimal investment strategy that maximizes the insurer’s expected terminal wealth and maintains the insurer’s solvency. If resource and time allow, we shall also study the precise large deviation for aggregate claims in a nonstandard renewal risk model in which claim sizes are negatively dependent with common consistently-varying-tailed distribution, and their arrival times are also negatively dependent. |
| List of Research Outputs |
| Dong Y., Wang G. and Yuen K.C., On the renewal risk model under a threhold strategy, Journal of Computational and Applied Mathematics. Elsevier, 2009, 230(1): 22-33. |
| Tang Q., Wang G. and Yuen K.C., Uniform tail asymptotics for the stochastic present value of aggregate claims in the renewal risk model, Insurance: Mathematics and Economics. Elsevier, 2010, 46(2): 362-370. |
| Tian G., Tang M.L., Yuen K.C. and Ng K.W., Further properties and new applications for the nested Dirichlet distribution, Computational Statistics and Data Analysis. Science Direct, 2010, 54(2): 394-405. |
| Wat K.P., Yuen K.C. and Li W.K., Time series insurance risk models with dependence structures (presented by Kam Pui Wat, PhD student), International Conference on Actuarial and Financial Risks, Shanghai, China. 2010. |
| Yuen K.C., Actuarial Analysis of some insurance risk models with dividend payments, First Conference in Statistical Finance, Singapore. 2009. |
| Yuen K.C., Computational Statistics and Data Analysis , Associate Editor , 2010. |
| Yuen K.C., On The Compound Binomial Risk Model With Time-correlated Claims And Randomized Dividend Policy, Department of Mathematics, Suzhou University, Suzhou, China. 2009. |
| Yuen K.C., Li J., Wat K.P. and Li W.K., On the compound binomial risk model with dividends and time-correlated claims (presented by Kam Pui Wat, PhD student), The 14th International Congress on Insurance: Mathematics and Economics, Toronto, Canada. 2010. |
| Researcher : Zang Y |
| List of Research Outputs |
| Zang Y., Fung T.W.K. and Zheng G., Asymptotic powers for matched trend tests and robust matched trend tests in case-control genetic association studies, Computational Statistics and Data Analysis. Elsevier B.V., 2010, 54: 65-77. |
| Zang Y., Fung T.W.K. and Zheng G., Simple algorithms to calculate asymptotic null distributions of robust tests in case-control genetic association studies in R, Journal of Statistical Software. The American Statistical Association, 2010, 33(8): 1-24. |
| Researcher : Zhao J |
| List of Research Outputs |
| Zhao J. and Yu P.L.H., A note on variational Bayesian factor analysis, Neural Networks. Elsevier Ltd., 2009, 22: 988-997. |
| Researcher : Zhou J |
| List of Research Outputs |
| Zhou J., Ding J., Fung T.W.K. and Lin S., Detection of parent-of-origin effects using general pedigree data, Genetic Epidemiology. Wiley-Liss, Inc, 2010, 34: 151-158. |