Dept of Statistics & Actuarial Science

DEPT OF STATISTICS & ACTUARIAL SCIENCE

Project Title:	Personnel surveys
Investigator(s):	Bacon-Shone JH
Department:	Social Sciences Research Ctr
Source(s) of Funding:	Other Funding Scheme
Start Date:	03/1991

Abstract:

To study both general and specific problems relating to personnel management in public and private organisations with a view of increasing management effectiveness and efficiency. Research designs will be adopted on individual project basis.

Project Title:	An exploratory assessment of willingness to pay for health care in Hong Kong
Investigator(s):	Bacon-Shone JH, McGhee SM
Department:	Social Sciences Research Ctr
Source(s) of Funding:	Health Services Research Fund - Full Grants
Start Date:	11/1996

Abstract:

To explore factors affecting WTP in Hong Kong health care using qualitative research; to investigate the advantages and disadvantages of using contingent valuation (CV) and conjoint analysis (CA) as tools for assessing WTP for Hong Kong health care; to use CA and CV to obtain initial estimates of WTP in Hong Kong health care.

Project Title:	Compositional data analysis with zeros and endmembers - an application to the Hong Kong household expenditure survey
Investigator(s):	Bacon-Shone JH
Department:	Social Sciences Research Ctr
Source(s) of Funding:	Competitive Earmarked Research Grants (CERG)
Start Date:	10/2003

Abstract:

To address two important methodological questions and to assist in answering an important policy question relevant to Hong Kong. - The first methodolocial question is the handling of zeros in compositional data, which has been identified as a major weakness of the key paradigm which uses log ratios (which are undefined for zeros). - The second question is how best to extend previous work on what is called the end methods problem by geologists, which is the problem of identifying which set of compositions has been mixed into a composite composition and what is the mixing process.

Project Title:	The UN business crime victim survey: a comparative study of the prevalence of crime against business in Chinese cities
Investigator(s):	Bacon-Shone JH, Broadhurst RG
Department:	Social Sciences Research Ctr
Source(s) of Funding:	Competitive Earmarked Research Grants (CERG)
Start Date:	01/2005

Abstract:

To estimate the prevalence and severity of crimes against business in Chinese cities of varying economic development, legal infrastructure and risks of conventional crime; to estimate the willingness of business to report crime and the reasons for reporting or not to police; to provide capacity to undertake policy relevant longitudinal research on changes in business crime by feasible and routine periodic surveillance of crimes against business in HK and selected Chinese cities; to estimate the risk of organized criminal activities against business and assess the effectiveness of countermeasures and the potential of innovations in crime prevention.

Project Title:	Postgraduate Students Conference/Seminar Grants 2006/2007
Investigator(s):	Bacon-Shone JH
Department:	Social Sciences Research Ctr
Source(s) of Funding:	Postgraduate Students Conference / Seminar Grants
Start Date:	07/2006

Abstract:
To invite 4 speakers to attend the 2006 Summer School on Research Methods.

Researcher : Bai Y

List of Research Outputs

Bai Y., Zhu Z.Y. and Fung T.W.K., Partial linear models for longitudinal data based on quadratic inference functions , Research Report. Department of Statistics and Actuarial Science, HKU, 2006, 439: 1-23.

Researcher : Chan WS

List of Research Outputs

Leung G.M., Tin Y.K. and Chan W.S., Hong Kong's health spending projections through 2033, Health policy. 2007, 81: 93-101.

Researcher : Chen A

Project Title:	Stochastic modelling of interacting branching systems
Investigator(s):	Chen A
Department:	Statistics & Actuarial Science
Source(s) of Funding:	Competitive Earmarked Research Grants (CERG)
Start Date:	08/2006

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.

List of Research Outputs

Chen A., Pollett P., Li J. and Zhang H., A remark on the uniqueness of weighted Markov branching processes, Journal of Applied Probability. Applied Probability Trust, 2007, 44(1): 279-283.

Chen A., Pollett P., Li J. and Zhang H., Markovian bulk-arriving queues with negative arrivals and state-dependent control (I): construction and equilibrium property , Research Report. Department of Statistics and Actuarial Science, HKU, 2006, 440: 1-27.

Chen A., Pollett P., Li J. and Zhang H., Markovian bulk-arriving queues with negative arrivals and state-dependent control (II): hitting time and busy period distributions , Research Report. Department of Statistics and Actuarial Science, HKU, 2006, 442: 1-21.

Li J. and Chen A., Decay parameter and quasi-stationary distributions for stopped M^X /M/1 queueing models , Research Report. Department of Statistics and Actuarial Science, HKU, 2006, 45: 1-17.

Li J. and Chen A., Existence, uniqueness and ergodicity of Markov branching processes with immigration and instantaneous resurrection , Research Report. Department of Statistics and Actuarial Science, HKU, 2006, 441: 1-22.

Researcher : Chen Y

List of Research Outputs

Chen Y., Yuen K.C. and Ng K.W., Asymptotics for ruin probabilities of a two-dimensional renewal risk model with heavy-tailed claims , Research Report. Department of Statistics and Actuarial Science, HKU, 2007, 452: 1-12.

Chen Y. and Ng K.W., The ruin probability of the renewal model with constant interest force and negatively dependent heavy-tailed claims, Insurance: Mathematics and Economics . Elsevier B.V., 2007, 40(3): 415-423.

Researcher : Cheung KY

List of Research Outputs

Cheung K.Y. and Lee S.M.S., Bootstrap variance estimation for Nadaraya quantile estimator , Research Report. Department of Statistics and Actuarial Science, HKU, 2006, 436: 1-23.

Researcher : Chung YK

List of Research Outputs

Fung T.W.K., Hu Y. and Chung Y.K., On statistical analysis of forensic DNA: Theory, methods and computer programs, Forensic Science International. The Netherlands, Elsevier B.V., 2006, 162: 17-23.

Researcher : Fong PW

List of Research Outputs

Fong P.W., Li W.K. and An H.Z., A simple multivariate ARCH model specified by random coefficients, Computational Statistics & Data Analysis. Holland, Elsevier B.V., 2006, 51(3): 1779-1802.

Researcher : Fung TWK

Project Title:	Individuality of handwritten numerals and characters in local population
Investigator(s):	Fung TWK
Department:	Statistics & Actuarial Science
Source(s) of Funding:	Small Project Funding
Start Date:	11/2004

Abstract:

To study if hierarchical cluster analysis is useful to group the quantified handwritten features of numerals and characters which are categorical in nature; to see if the quantified features are statistically independent of one another; the evaluation of the probability of occurrence of certain features would be simplified if the independence condition is satisfied; to investigate if the hypothesis of individuality of handwritten numerals and characters in local population is valid.

Project Title:	The Transmission Disequilibrium Test in the Presence of Genomic Imprinting
Investigator(s):	Fung TWK
Department:	Statistics & Actuarial Science
Source(s) of Funding:	Seed Funding Programme for Basic Research
Start Date:	01/2006

Abstract:

Key Issues, Background and Problems Being Addressed: The transmission/disequilibrium test (TDT) is a powerful and major approach to search for genes underlying human complex/common diseases. It was introduced originally by Spielman et al. (1993) based on the case-parents trios, to test directly for linkage between the marker locus and a disease susceptibility locus (DSL) when association due to linkage disequilibrium (LD) is present. When both parents' marker genotypes were unavailable, Ewens and Spielman (1995) extended the TDT to the S-TDT for use in sibships with at least one affected individual and one unaffected individual. When only one of the parents was available, Sun et al. (1999) proposed a test, termed the 1-TDT, for use with marker genotypes of the affected individuals and the available parents. A number of investigators worked on the statistical power of the TDT. For example, Xiong and Guo (1998) demonstrated that the power was a function of several genetic parameters including the recombination fraction, penetrance, age of mutant disease allele, marker allele frequency, recurrent mutation rates at the marker and/or disease locus, and initial linkage disequilibrium. Knapp (1999) presented a rigorous method for obtaining the power of the TDT for samples consisting of families with either a single affected child or affected sib pairs, based on the asymptotic normality of the maximum likelihood estimator. Knapp (1999) also pointed out that the approach proposed by Camp (1997) on sample size calculations was erroneous. Wang and Sun (2000) presented a general method to obtain the number of families required to detect linkage disequilibrium and compared the relative power of the TDT, S-TDT and 1-TDT. Deng and Chen (2001) developed a simple and accurate analytical method for computing the statistical power of various TDT, incorporating unrelated cases and controls and their parents into a single analysis therein. Iles (2002) demonstrated a basic model that allowed analytical comparisons of the three methods of Risch and Merikangas (1996), McGinnis (1998, 2000), Tu and Whittemore (1999) and Knapp (1999), for calculating the power of TDT. Genomic imprinting, also known as “parent-of-origin effect”, is an important epigenetic factor. More than 1% of all mammalian genes are known to be imprinted (Morison et al. 2001). Morison et al. (2001) had constructed an imprinted-gene database which contained 230 records in humans at the time of submission (URL: www.otago.ac.nz/IGC). The transcriptional activity of an imprinted gene is dependent on the parent-of-origin of the allele. Imprinting can happen when either the paternal or the maternal copy of a gene is inactivated. DNA methylation and the differential packing density of DNA by histone proteins are two mechanisms known to be involved in the process of imprinting, which is determined by chromosomal region of DNA or by differences in chromatin structure (Hall 1990; Ainscough and Surani 1996, Bartolomei and Tilghman 1997; Strauch et al. 2000; Knapp and Strauch 2004). Reviews of mechanisms and function of genomic imprinting can be referred to Pfeifer (2000), Reik and Walter (2001) and Wilkins and Haig (2003). Parent-of-origin effects have been demonstrated in several genetic disorders, such as Beckwith-Wiedemann, Prader-Willi, and Angleman syndromes (Falls et al. 1999). The statistical power of tests for genetic linkage would be affected when the genes are imprinted. Much work has been done recently to incorporate parent-of-origin effects into existing tests for linkage. Weinberg et al. (1998) investigated the effects in genetic studies of case-parents trios by a log-linear approach. Weinberg (1999) proposed the methods for detection of parent-of-origin. Both of them introduced some parameters to describe the parental effects. Knapp and Strauch (2004) derived the asymptotic distribution of the likelihood-ratio test, allowing for imprinting under the null hypothesis of no linkage, by extending Holmans’ (1993) possible triangle test for affected sib pairs. Wu et al. (2005) proposed a robust generalized minimax test for linkage, based on alleles that were shared identical by descent by different affected sib pairs, allowing for parent-of-origin effects and sex-specific recombination rates. In the presence of genomic imprinting, the methods for incorporating of parent-of-origin into linkage analysis of quantitative traits can be referred to the work of Hanson et al. (2001), Shete and Amos (2002), Shete et al. (2003), among some others. Objectives: According to a rule by Haldane (1922), it is more common to have crossovers in the homogametic sex (e.g., XX) than in the heterogametic sex (e.g., XY) (Ott 1999). In other words, recombination rates are sex-specific. The recombination rate for human females is on the average 60% higher than that for human males (Fann and Ott 1995; Broman et al. 1998), but the rates can be highly variable in some regions of chromosome (Wu et al. 2005). In linkage analysis, sex-specific recombination rates will be a consequence of imprinting and Smalley (1993) suggested the utilization of this information for possible identification of traits undergoing imprinting. The recombination fractions are sex-specific and imprinting is known or believed to play an important role in many genetically complex traits, such as type I diabetes, polycystic ovarian syndrome (Bennett et al. 1997), atopy, celiac disease, cancer, epilepsy, or bipolar disorder (Knapp and Strauch 2004). We consider the following objectives in this project: (i) To study if the statistical power of TDT will change when there is genomic imprinting, i.e. how does the parent-of-origin affect the power of the TDT and is the TDT still a valid test for linkage when imprinting is present? (ii) To construct a valid test (or tests) for imprinting, based on the case-parents trios. (iii) To investigate if we can find a new method to improve the TDT for more powerful test for linkage. (iv) To see the role of sex-specific recombination fractions on the above problems. (Details of the reference list are omitted for brevity.)

Project Title:	Generalized estimating equations for structural parameters in regression credibility models of actuarial science
Investigator(s):	Fung TWK, Zhu ZY
Department:	Statistics & Actuarial Science
Source(s) of Funding:	Competitive Earmarked Research Grants (CERG)
Start Date:	07/2006

Abstract:

We propose to use the GEE approach for estimation of the structural parameters of a regression credibility model. Specifically, we intend (1) to construct estimating equations for regression credibility models using the Bühlmann, Bühlmann-Straub and Hachemeister frameworks; (2) to investigate the performance of various GEE estimators; (3) to study the impact on efficiency and robustness of GEE estimators for different choices of working covariance matrix in the GEE framework; (4) to ascertain the more robust GEE approach to be a better surrogate in most scenarios for the Hachemeister method which has been criticized for frequently generating invalid estimates; (5) to reparameterize the GEE formulation and transform the regression credibility model to solve the numerical computation problem commonly found in the estimation routine, while maintaining the efficiency and invariance of GEE estimators.

List of Research Outputs

Bai Y., Zhu Z.Y. and Fung T.W.K., Partial linear models for longitudinal data based on quadratic inference functions , Research Report. Department of Statistics and Actuarial Science, HKU, 2006, 439: 1-23.

Fung T.W.K. and Yu P.L.H., New-based learning in statistics, Proceedings of the 7th International Conference on Teaching Statistics. 2006.

Fung T.W.K., Hu Y. and Chung Y.K., On statistical analysis of forensic DNA: Theory, methods and computer programs, Forensic Science International. The Netherlands, Elsevier B.V., 2006, 162: 17-23.

Fung T.W.K., Panelist, "Panel discussions on promoting Pan-Asian collaboration between statisticians", International Association for Statistical Computing Asian Regional Section Special Conference, Seoul, Korea, June 2007.

Fung T.W.K., Robust inference in generalized partial linear models for longitudinal data, COMPSTAT 2006: 17th Symposium of IASC on Computational Statistics, held in Rome, August 2006.

Fung T.W.K., Transmission disequilibrium tests and imprinting effects tests based on case-parents trios and case-parent pairs, Colloquium held at Hong Kong Baptist University, Faculty of Science, Department of Mathematics, Hong Kong, 27 April, 2007.

Fung T.W.K., Transmission disequilibrium tests and imprinting effects tests based on case-parents trios and case-parent pairs, Seminar held at The Hong Kong Polytechnic University, Department of Applied Mathematics, Hong Kong, 11 May, 2007.

Hu Y., Zhou J. and Fung T.W.K., An extension of the transmission disequilibrium test incorporating imprinting, Genetics. USA, Genetics Society of America, 2007, 175: 1489-1504.

Hu Y., Zhou J., Sun F.Z. and Fung T.W.K., The transmission disequilibrium test and imprinting effects test based on case-parent pairs, Genetic Epidemiology. USA, Wiley-Liss, Inc., 2007, 31: 273-287.

Li C.K., Yang C.T., Poon N.L. and Fung T.W.K., Significance of sequence of strokes in Chinese handwriting examination, Journal of Forensic Sciences. USA, American Academy of Forensic Sciences, 2007, 52(2): 467-472.

Lo C.H., Fung T.W.K. and Zhu Z.Y., Generalized estimating equations for variance and covariance parameters in regression credibility models, Insurance Mathematics and Economics. The Netherlands, Elsevier B.V., 2006, 39: 99-113.

Nyangoma S.O., Fung T.W.K. and Jansen R.C., Identifying influential multinomial observations by perturbation, Computational Statistics & Data Analysis. The Netherlands, Elsevier, B.V., 2006, 50: 2799-2821.

Zhou J., Hu Y. and Fung T.W.K., A simple method for detection of imprinting effects based on case-parents trios, Heredity. UK, Nature Publishing Group, 2007, 98: 85-91.

Researcher : Hu Y

List of Research Outputs

Fung T.W.K., Hu Y. and Chung Y.K., On statistical analysis of forensic DNA: Theory, methods and computer programs, Forensic Science International. The Netherlands, Elsevier B.V., 2006, 162: 17-23.

Hu Y., Zhou J. and Fung T.W.K., An extension of the transmission disequilibrium test incorporating imprinting, Genetics. USA, Genetics Society of America, 2007, 175: 1489-1504.

Hu Y., Zhou J., Sun F.Z. and Fung T.W.K., The transmission disequilibrium test and imprinting effects test based on case-parent pairs, Genetic Epidemiology. USA, Wiley-Liss, Inc., 2007, 31: 273-287.

Zhou J., Hu Y. and Fung T.W.K., A simple method for detection of imprinting effects based on case-parents trios, Heredity. UK, Nature Publishing Group, 2007, 98: 85-91.

Researcher : Jin S

List of Research Outputs

Jin S. and Li W.K., Modeling panel time series with mixture autoregressive model, Journal of Data Science. Taiwan, Academia Sinica, 2006, 4(4): 425-446.

Researcher : Kwok SSM

List of Research Outputs

Kwok S.S.M. and Li W.K., A note on diagnostic checking of the double autoregressive model , Research Report. Department of Statistics and Actuarial Science, HKU, 2007, 447: 1-21.

Researcher : Lai PY

List of Research Outputs

Lee S.M.S. and Lai P.Y., Improving coverage accuracy of block bootstrap confidence intervals , Research Report. Department of Statistics and Actuarial Science, HKU, 2006, 435: 1-44.

Researcher : Lam KF

Project Title:	Semiparametric regression analysis for clustered interval censored survival data
Investigator(s):	Lam KF
Department:	Statistics & Actuarial Science
Source(s) of Funding:	Competitive Earmarked Research Grants (CERG)
Start Date:	01/2006

Abstract:

To propose an asymptotically semiparametric efficient estimator for the regression parameter in the conditional proportional hazards model, at which the effect of the associated covariate X is assumed to be linearly related to log of the hazard function and the dependence parameter with clustered interval censored data. The rate of convergence, the asymptotic properties of the estimator will be studied; to provide a simple efficient estimation method to estimate the parameters of the semiparametric random effects proportional hazards model with a partially linear relative risk function that combines the linear part based on the usual parametric regression models and a nonlinear, nonparametric regression part to explore the nature of relationships between some covariates and the failure time.

List of Research Outputs

Chan A.O.O., Chu K.M., Huang C.Y., Lam K.F., Leung S.Y., Sun Y., Ko S., Xia H.H.X., Cho C.H., Hui W.M., Lam S.K. and Rashid A., Association between Helicobacter pylori infection and interleukin 1beta polymorphism predispose to CpG island methylation in gastric cancer, GUT. 2007, 56: 595-597.

Chan A.O.O., Hui W.M., Lam K.F., Leung G.K.L., Yuen R.M.F., Lam S.K. and Wong B.C.Y., Familial aggregation in constipated subjects in a tertiary referral center, American Journal of Gastroenterology. Blackwell Publishing, 2007, 102: 149-152.

Chan A.O.O., Lam K.F., Hui W.M., Leung G.K.L., Wong Y.H., Lam S.K. and Wong B.C.Y., Influence of Positive Family History on Clinical Characteristics of Functional Constipation, Clinical Gastroenterology and Hepatology. Elsevier, 2007, 5(2): 197-200.

Chan A.O.O., Hui W.M., Leung G.K.L., Tong T.P.H., Hung I.F.N., Chan P., Hsu A., But D., Wong B.C.Y., Lam S.K. and Lam K.F., Patients with functional constipation do not have increase prevalence of colorectal cancer precursors”, GUT. British Medicial Journal, 2007, 56: 451-452.

Chan O.O., Lam K.F., Tong T.P.H., Siu D.C.W., Jim M.H., Hui W.M., Lai K.C., Yuen R.M.F., Lam S.K. and Wong B.C.Y., Coexistence between colorectal cancer/adenoma and coronary artery disease: results from 1382 patients. , Alimentary Pharmacology and Therapeutics. 2006, 24(3): 535-9.

Cheung T.K., Lam K.F., Hu H.C., Wong W.M., Hui W.M., Lam C.L.K., Lai K.C., Lam S.K. and Wong B.C.Y., Positive association between gastro-oesophageal reflux disease and irritable bowel syndrome in a Chinese population, Alimentary Pharmacology & Therapeutics. Blackwell Publishing, 2007, 25: 1099-1104.

Cheung Y.B. and Lam K.F., Bivariate Poisson-Poisson model of zero-inflated absenteeism data, Statistics in Medicine. John Wiley & Sons, Ltd., 2006, 25: 3707-3717.

Dickman R., Bautista J.M., Wong W.M., Bhatt R., Beeler J.N., Malagon I., Risner-Adler S., Lam K.F. and Fass R., Comparison of esophageal acid exposure distribution along the esophagus among the different gastroesophageal reflux disease (GERD) groups, American Journal of Gastroenterology. Blackwell Publishing, 2006, 101: 2463-2469.

Hu H.C., Lam S.K., Lam C.L.K., Wong R.W.M., Lam K.F., Lai K.C., Wong Y.H., Wong B.C.Y., Chan O.O., Chan C.K., Leung G.M. and Hui W.M., Comparison between Empirical Prokinetics, Helicobacter Test-and-Treat and Empirical Endoscopy in Primary-Care Patients Presenting with Dyspepsia: A One-Year Study, World Journal of Gastroenterology. 2006, 12(31): 5010-5016.

Lam K.F., Xue H.Q. and Cheung Y.B., Semiparametric analysis of zero-inflated count data, Biometrics. The International Biometric Society, 2006, 62: 996-1003.

Lam K.F., Statistical Advisor for The Journal of Orthopaedic Surgery . Hong Kong Academy of Medicine Press, 2007.

Lam K.F., In: Hong Kong Medical Journal, Hong Kong Academy of Medicine & HK Medical Association, 2007.

Wong M.C.M., Lam K.F. and Lo E.C.M., Faculty Research Output Prize, 2006.

Xue H.Q., Lam K.F., Cowling B.J. and de Wolf F., Semi-parametric accelerated failure time regression analysis with application to interval censored HIV/AIDS data, Statistics in Medicine. John Wiley & Sons, Ltd., 2006, 25: 3850-3863.

Yip P.S.F., Hsieh Y.H., Xu Y., Lam K.F., King C.C. and Chang H.L., Assessment of intervention measures for the 2003 SARS epidemic in Taiwan by use of a back-projection method, Infection Control and Hospital Epidemiology . University of Chicago Press, 2007, 28(5): 525-530.

Researcher : Lam Y

List of Research Outputs

Chen J., Li K.H. and Lam Y., Bayesian single and double variable sampling plans for the Weibull distribution with censoring, European Journal of Operational Research. The Netherlands, Elsevier B.V., 2007, 177(2): 1062-1073.

Lam Y., A geometric process warranty model-game theory approach , Research Report. Department of Statistics and Actuarial Science, HKU, 2007, 450: 1-22.

Lam Y., A threshold geometric process maintenance model, Research Report. Department of Statistics and Actuarial Science, HKU, 2007, 449: 1-16.

Lam Y., Li K.H., Ip W.C. and Wong H., Sequential variable sampling plan for normal distribution , European Journal of Operational Research . The Netherlands, Elsevier B.V., 2006, 172(1): 127-145.

Lam Y. and Tang Y.Y., The analytic solution of an integral equation in geometric process , Research Report. Department of Statistics and Actuarial Science, HKU, 2007, 451: 1-32.

Researcher : Lee SMS

Project Title:	Variance estimation for kernel-smoothed sample quantiles
Investigator(s):	Lee SMS
Department:	Statistics & Actuarial Science
Source(s) of Funding:	Small Project Funding
Start Date:	12/2006

Abstract:

To the problem of estimating a population quantile based on a random sample, the sample quantile provides a natural solution but suffers from a lack of efficiency due to the variability of individual order statistics. The kernel-smoothed sample quantile, defined as the inverse of the kernel-smoothed empirical distribution of the observed sample, succeeds in reducing variability and is hence a more efficient estimator. Unlike the case of sample quantile, the important problem of variance estimation for the kernel-smoothed sample quantile has, however, received little attention in the literature. The proposed project investigates the properties and explores ways to improve the accuracy of the bootstrap procedure for estimating the variance of the kernel-smoothed sample quantile. Specifically, the project has the following objectives: (1) establish an asymptotic expansion for the variance of the kernel-smoothed sample quantile, against which its bootstrap estimators can be compared and evaluated asymptotically; (2) establish the order of the relative error of the ordinary bootstrap variance estimator, upon which modifications of the bootstrap are sought to improve; (3) investigate the effects of smoothing the bootstrap by the kernel method on the relative error of the variance estimator and hence derive the optimal choice of bandwidth for smoothing the bootstrap; (4) investigate the effects of calibrating the order of the kernel-smoothed bootstrap sample quantiles on the relative error of the variance estimator and determine the precise adjustment made to the quantile order for maximum improvement; (5) develop a practical computational procedure for optimal calibration of the quantile order as determined by objective (4); (6) conduct a simulation study on empirical performance of the various bootstrap procedures developed in the project.

Project Title:	An omnibus procedure for robust depth-based confidence regions
Investigator(s):	Lee SMS
Department:	Statistics & Actuarial Science
Source(s) of Funding:	Competitive Earmarked Research Grants (CERG)
Start Date:	01/2007

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.

List of Research Outputs

Chan K.Y.F., Lee S.M.S. and Ng K.W., Minimum variance unbiased estimation based on bootstrap iterations, Statistics and Computing. 2006, 16: 267-277.

Cheung K.Y. and Lee S.M.S., Bootstrap variance estimation for Nadaraya quantile estimator , Research Report. Department of Statistics and Actuarial Science, HKU, 2006, 436: 1-23.

Ho H.S. and Lee S.M.S., Iterated bootstrap-t confidence intervals for density functions, International Conference on Ordered Statistical Data and Inequalities: Theory and Applications. Amman, Jordan, 2007.

Lee S.M.S. and Lai P.Y., Improving coverage accuracy of block bootstrap confidence intervals , Research Report. Department of Statistics and Actuarial Science, HKU, 2006, 435: 1-44.

Lee S.M.S. and Pun M.C., On m out of n bootstrapping for nonstandard M-estimation with nuisance parameters, Journal of the American Statistical Association. 2006, 101: 1185-1197.

Lok W.S. and Lee S.M.S., Data depth for high-dimensional data, International Conference on Ordered Statistical Data and Inequalities: Theory and Applications. Amman, Jordan, 2007.

Lok W.S. and Lee S.M.S., Two new statistical depth functions for high-dimensional data, COMPSTAT 2006, 17th Symposium of IASC-ERS. 2006.

Researcher : Li G

List of Research Outputs

Li G. and Li W.K., Least absolute deviation estimation for unit root processes with GARCH errors , Research Report. Statistics and Actuarial Science, HKU, 2006, 444: 1-40.

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:	Competitive Earmarked Research Grants (CERG)
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.

List of Research Outputs

Ching W.K., Siu K.T.K., Fung S.L., Ng K.P. and Li W.K., Interactive Hidden Markov Models and Their Applications, IMA Journal of Management Mathematics. Oxford University Press, 2007, 18: 85-97.

Fong P.W., Li W.K. and An H.Z., A simple multivariate ARCH model specified by random coefficients, Computational Statistics & Data Analysis. Holland, Elsevier B.V., 2006, 51(3): 1779-1802.

Fong P.W., Li W.K., Yau C.W. and Wong C.S., On a Mixture Vector Autoregressive Model, The Canadian Journal of Statistics. 2007, 35: 135-150.

Jin S. and Li W.K., Modeling panel time series with mixture autoregressive model, Journal of Data Science. Taiwan, Academia Sinica, 2006, 4(4): 425-446.

Kwok S.S.M. and Li W.K., A note on diagnostic checking of the double autoregressive model , Research Report. Department of Statistics and Actuarial Science, HKU, 2007, 447: 1-21.

Kwok W.H.K.P., Leung K.M.Y., Lui C.S.G., Li W.K., Lam P.K.S. and Lee J.H.W., Derivation of better ammonia water quality criteria for saltwater, Society of Environmental Toxicology and Chemistry (SETAC) Asia/Pacific 2006 meeting, 18-20 September, 2006, Peking University, Beijing, China. 2006.

Li G. and Li W.K., Least absolute deviation estimation for unit root processes with GARCH errors , Research Report. Statistics and Actuarial Science, HKU, 2006, 444: 1-40.

Li W.K., Fellow, Institute of Mathematical Statistics, 2006.

Lui G.C.S., Li W.K., Leung K.M.Y., Lee J.H.W. and Jayawardena A.W., Modelling algal blooms using vector autoregressive model with exogeneous variables and long memory filter, Ecological Modelling. 2007, 200: 130-138.

Wan L.M., Yuen K.C. and Li W.K., Analysis of an insurance risk model with thinning dependence and common shock, Journal of Actuarial Practice. U.S.A., Absalom Press, 2006, 13: 147-164.

Wong S.T. and Li W.K., A note on the estimation of extreme value distributions using maximum product of spacings, Institute of Mathematical Statistics. U.S.A., Institute of Mathematical Statistics, 2006, 52: 272-283.

Wu H., Yu P.L.H. and Li W.K., Value at risk estimation using independent component analysis-generalized autoregressive conditional heteroscedasticity (ICA-GARCH) models, International Journal of Neural Systems. Singapore, World Scientific, 2006, 16(5): 371-382.

Xia Y., Li W.K. and Tong H., Threshold variable selection using nonparametric methods, Statistica Sinica. Taipei, Taiwan, Institute of Statistical Science, Academia Sinica, 2007, 17(1): 265-287.

Yuen K.C., Wang G. and Li W.K., The Gerber-Shiu expected discounted penalty function for risk processes with interest and a constant dividend barrier, Insurance: Mathematics and Economics. 2007, 40(1): 104-112.

Zhang Z., Li W.K. and Yuen K.C., On a mixture GARCH time series model, Journal of Time Series Analysis. Blackwell, 2006, 27(4): 577-597.

Zhou X. and Li W.K., Modeling threshold conditional heteroscedasticity with regime-dependent skewness and kurtosis , Research Report. Department of Statistics and Actuarial Science, HKU, 2007, 448: 1-28.

Researcher : Lo CH

List of Research Outputs

Lo C.H., Fung T.W.K. and Zhu Z.Y., Generalized estimating equations for variance and covariance parameters in regression credibility models, Insurance Mathematics and Economics. The Netherlands, Elsevier B.V., 2006, 39: 99-113.

Researcher : Lok WS

List of Research Outputs

Lok W.S. and Lee S.M.S., Data depth for high-dimensional data, International Conference on Ordered Statistical Data and Inequalities: Theory and Applications. Amman, Jordan, 2007.

Lok W.S. and Lee S.M.S., Two new statistical depth functions for high-dimensional data, COMPSTAT 2006, 17th Symposium of IASC-ERS. 2006.

Researcher : Lui CSG

List of Research Outputs

Kwok W.H.K.P., Leung K.M.Y., Lui C.S.G., Li W.K., Lam P.K.S. and Lee J.H.W., Derivation of better ammonia water quality criteria for saltwater, Society of Environmental Toxicology and Chemistry (SETAC) Asia/Pacific 2006 meeting, 18-20 September, 2006, Peking University, Beijing, China. 2006.

Kwok W.H.K.P., Bjørgesæter A., Leung K.M.Y., Lui C.S.G., Gray J.S., Shin P.K.S. and Lam P.K.S., Deriving site-specific sediment quality guidelines from field-based species sensitivity distributions: Hong Kong case study, Society of Environmental Toxicology and Chemistry (SETAC) Asia/Pacific 2006 meeting, 18-20 September, 2006, Peking University, Beijing, China. 2006.

Leung K.M.Y., Kwok W.H.K.P., Lui C.S.G., Chu V.K.H., Lam P.K.S., Morritt D., Maltby L., Brock T.C.M., Brink P.J.V. and Warne M.S.J., Comparison of tropical and temperate freshwater species acute sensitivities to chemicals: Implications for deriving safe extrapolation factors, Society of Environmental Toxicology and Chemistry (SETAC) Asia/Pacific 2006 meeting, 18-20 September, 2006, Peking University, Beijing, China. 2006.

Researcher : Lui GCS

List of Research Outputs

Kwok K.P.W.H., Leung K.M.Y., Lui G.C.S., Chu V.K.H., Lam P.K.S., Morritt D., Maltby L., Brock T.C.M., Van Den Brink P.J., Warne M.S.T.J. and Crane M., Comparison of tropical and temperate freshwater animal species' acute sensitivity to chemicals: implication for deriving safe extrapolation factor, Integrated Environmental Assessment and Management. SETAC, 2007, 3: 49-67.

Lui G.C.S., Li W.K., Leung K.M.Y., Lee J.H.W. and Jayawardena A.W., Modelling algal blooms using vector autoregressive model with exogeneous variables and long memory filter, Ecological Modelling. 2007, 200: 130-138.

Researcher : Ng KW

List of Research Outputs

Chan K.Y.F., Lee S.M.S. and Ng K.W., Minimum variance unbiased estimation based on bootstrap iterations, Statistics and Computing. 2006, 16: 267-277.

Chen Y., Yuen K.C. and Ng K.W., Asymptotics for ruin probabilities of a two-dimensional renewal risk model with heavy-tailed claims , Research Report. Department of Statistics and Actuarial Science, HKU, 2007, 452: 1-12.

Chen Y. and Ng K.W., The ruin probability of the renewal model with constant interest force and negatively dependent heavy-tailed claims, Insurance: Mathematics and Economics . Elsevier B.V., 2007, 40(3): 415-423.

Li Z., Ng K.W., Tan K.S. and Yang H., Best CRP investment strategies for Dynamic Portfolio Selection, International Journal of Theoretical and Applied Finance. 2006, 9: 951-966.

Li Z.F., Ng K.W. and Tan K.S., Best CRP investment strategies for Dynamic Portfolio Selection, International Journal of Theoretical and Applied Finance. 2006, 9: 1-16.

Ng K.W., Yang H. and Zhang L., Upper Bounds for Ruin Probability in a filtered compound Poisson model , International Journal on Statistics and Systems. 2006, 1: 191-201.

Tang M.L., Ng K.W., Tian G.L. and Tan M., On improved EM algorithm and confidence interval construction for incomplete r´c tables , Computational Statistics & Data Analysis . Elsevier B.V., 2007, 51(6): 2919-2933.

Tian G.L., Tan M. and Ng K.W., An exact non-iterative sampling procedure for discrete missing data problems, Statistica Neerlandica. Blackwell Publishing, 2007, 61(2): 232-242.

Researcher : Pun MC

List of Research Outputs

Lee S.M.S. and Pun M.C., On m out of n bootstrapping for nonstandard M-estimation with nuisance parameters, Journal of the American Statistical Association. 2006, 101: 1185-1197.

Researcher : Siu KBB

List of Research Outputs

Siu K.B.B. and Yang H., Expected shortfall under a model with market and credit risks, In: R. Mamon and R. Elliott, Hidden Markov Models in Finance. Springer, 2007, 91-100.

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
Researcher : Wan LM

List of Research Outputs

Wan L.M., Yuen K.C. and Li W.K., Analysis of an insurance risk model with thinning dependence and common shock, Journal of Actuarial Practice. U.S.A., Absalom Press, 2006, 13: 147-164.

Researcher : Wang Q

Project Title:	Statistical Analysis Based on Complex Data
Investigator(s):	Wang Q
Department:	Statistics & Actuarial Science
Source(s) of Funding:	Seed Funding Programme for Basic Research
Start Date:	03/2006

Abstract:

In this project I will develop statistical inference tools and approaches with complex data including surrogate data and validation data, missing data and censored data. The project has the following objectives. 1) I will develop estimation approaches for a probability density and a regression function with surrogate data and validation sampling. A regression calibration kernel approach will be applied to estimation of the probability density and regression function to incorporate the information contained in both surrogate variables and validation sample. I intend to define two weighted probability density estimators which may have less asymptotic variances than the regression calibration kernel density estimator. I intend to prove that the proposed estimators are asymptotically normal, and derive the asymptotic representations for the mean squared error and mean integrated square error of the proposed estimators, respectively. 2) Let Y be the response variable and X the covariables. Let f(y|x,beta) be the conditional probability density of Y, which is parameterized up to an unknown parameter vector \beta. For making inference on \beta, it is typically assumed that (X,Y) are available. In many applications, however, the exact measurement of X may be difficult, time consuming or expensive, so surrogate {\wt X} is observed instead of X. Generally, the relation between the surrogate {\wt X} and the true variable X can be rather complicated. The usual additive error model may not be sufficient to describe this relation. In this project, I intend to consider settings where some validation data are available to relate X and {\wt X} without specifying any error structure model. A kernel-based likelihood scoring methods with the help of validation data will be developed to define the estimator of $\beta$. Also, I will develop a semiparametric likelihood method by combining parametric likelihood and empirical likelihood to incorporate auxiliary information for improving inference. 3) In this project, I intend to consider a semiparametric model which parameterizes only the relation of interest, f(y|x,\theta), and allows the marginal distribution G(.) of X to be completely arbitrary. I intend to define the estimator of \theta. We use multiple imputation method to impute missing responses and make inference on \theta by semi-empirical likelihood approaches. 4) Consider the partial linear model with the covariables missing at random and response variables are randomly right censored. A model calibration approach and a weighting approach will be developed to define the estimators of the parametric part and nonparametric part in the partial linear model, respectively. I intend to prove that the estimators for the parametric part are asymptotically normal and the estimators of g(.) convergent to g(.) with an optimal convergent rate. 5) I will develop methods for estimating a survival function and a hazard function with censoring indicators missing at random. e.g., we will define several asymptotically efficient PL estimators. I intend to prove all the estimators to be strong uniformly consistent, weakly converge to a Gaussian process and be asymptotically efficient.

Project Title:	Statistical analysis with missing survival data at random
Investigator(s):	Wang Q
Department:	Statistics & Actuarial Science
Source(s) of Funding:	Competitive Earmarked Research Grants (CERG)
Start Date:	09/2006

Abstract:

(1) Missing survival data arises in a number of applied fields such as medical study,biology,public health, epidemiology,clinic trial and so on. Statistical inference tools and approaches for various models and statistical characteristics will be developed under random censorship when some data are missing at random. We consider partial linear models, semiparametric transformation models, proportional hazard models with a partial linear link function when covariables are missing at random. Also, we consider the problem of estimating a survival function, hazard function and parameters in semiparametric transformation models when censoring indicators are missing at random. The project has the following objectives. (2) 1.1) Consider the partial linear model with the covariables missing at random when response variables are censored. A model calibration approach and a weighting approach will be developed to define the estimators of the parametric part and nonparametric part in the partial linear model,respectively. I intend to prove that the estimators for the parametric part are asymptotically normal and the estimators of nonparametric part converge with an optimal convergent rate. It is worth noting that linear or semiparametric partial linear models are useful alternatives to the commonly used Cox model when the proportional hazards assumption is not valid. However, no method is available in the literature for checking the linearity assumption of a linear model with right censored data. The partial linear model is a natural compromise between the linear model and the fully nonparametric model. It allows only some of the predictors to be modeled linearly, with others being modeled nonparametrically. It is a dimension reduction model compared to the fully nonparametric model. In practice, the partially linear model has been applied to many fields such as biometrics, medical study, economics and so on. Hence, the work will be of practical significance.(3)1.2). Semiparametric transformation models will be considered when the covariables are missing at random and the life variables are censored. An inverse probability weighted estimating equation approach will be developed for estimating of regression parameters.The estimation procedure of the survival probability at given covariate level also will be provided. Asymptotic properties for the suggested estimators will be investigated. This includes studying the consistency of the estimators and asymptotic distribution. This class of regression models includes the proportional hazards and proportional odds models as two special examples. The research can be applied to medical study and clinic trails. (4)1.3) Consider the extended proportional hazard models with a partial linear function When some components of covariable are missing at random, I intend to develop some methods for estimating the parameter vector and nonparametric part under random censorship. One of these approaches that I will develop is a local-likelihood score equation method by combining a weighted local partial likelihood and a score approach to define the estimators of the parameter vector and the nonparametric part. I intend to investigate their asymptotic properties including strong consistency and asymptotic normality etc. This model is an important extension of the Cox model and the proportional hazard model considered by Fan and Kong (1997, Ann. Statist.). The proposed procedures for this model can be applied to study of some practical problems such as survival analysis, medical study and biometrics. The work will have a practical impact. (5) 1.4) I will develop methods for estimating a survival function and a hazard function with censoring indicators missing at random. e.g., we will define several asymptotically efficient PL estimators. I intend to prove all the estimators to be strongly uniformly consistent, weakly converge to a Gaussian process and be asymptotically efficient. (6) 1.5) I will develop a calibrated estimation equation approach for estimating the parameters in the semiparametric transformation models with censoring indicators missing at random. I intend to give the asymptotic distribution of the estimators.(7) The project deals with important statistical inference problems in survival analysis when data are missing at random. This work will make original contributions to the foundations, theory development and methodology in survival analysis. If successful, it would have a significant and broad impact on the theory, methodology and application of survival analysis, and would be of great theoretical and practical interest.

List of Research Outputs

Wang L. and Wang Q., Empirical bayes estimation under random censorship, Southeast Asian Bulletin of Mathematics. 2006, 30(4): 779-788.

Wang Q. and Sun Z., Estimation In Partially Linear Models With Missing Responses At Random, Journal Of Multivariate Analysis. USA, Elsevier, 2007, 98: 1470-1493.

Wang Q. and Shen J., Estimation and confidence bands of a conditional survival function with censoring indicators missing at random , Research Report. Department of Statistics and Actuarial Science, HKU, 2006, 438: 1-26.

Wang Q., Hazard Function Estimation With Censoring Indicators Missing At Random, The 5th Statistical Conference of Mainland China and Taiwan. 2006.

Wang Q., Identifying Coexpressed Genes , In: Wolfgang Hardle, Yuichi Mori and Philippe Vieu, Statistical Methods For Biostatistics And Related Fields. Germany, Springer, 2006, 125-146.

Wang Q. and Yu K., Likelihood-based Kernel Estimation In Semiparametric Errors-in-covariables Models With Validation Data, Journal of Multivariate Analysis. USA, Elsevier, 2007, 98: 455-480.

Wang Q., M-estimators Based On Inverse Probability Weighted Estimating Equations With Response Missing At Random, Communications In Statistics--theory And Methods. USA, Taylor & Francis, 2007, 36: 1091-1103.

Wang Q., Probability density estimation with cause-of-death data missing at random , Research Report. Department of Statistics and Actuarial Science, HKU, 2007, 446: 1-30.

Wang Q., Probability density estimation with data missing at random when covariables are present , Research Report. Department of Statistics and Actuarial Science, HKU, 2006, 443: 1-29.

Wang Q., Qihua Wang (Associate Editor), Science in China. 2007.

Wang Q., Wang Qihua (Associate Editor), Acta Appl. Math. Sinica. 2007.

Wang Q., Wang Qihua (Associate Editor), Journal of Applied Probability and Statistics. 2007.

Researcher : Wong ST

List of Research Outputs

Wong S.T. and Li W.K., A note on the estimation of extreme value distributions using maximum product of spacings, Institute of Mathematical Statistics. U.S.A., Institute of Mathematical Statistics, 2006, 52: 272-283.

Researcher : Wu H

List of Research Outputs

Wu H., Yu P.L.H. and Li W.K., Value at risk estimation using independent component analysis-generalized autoregressive conditional heteroscedasticity (ICA-GARCH) models, International Journal of Neural Systems. Singapore, World Scientific, 2006, 16(5): 371-382.

Researcher : Wu HC

List of Research Outputs

Wu H.C. and Yu P.L.H., ICLUS: A robust and scalable clustering model for time series via independent component analysis, In: Peter Fleming, International Journal of Systems Science. UK, Taylor & Francis, 2006, 37(13): 987-1001.

Researcher : Xi L

List of Research Outputs

Xi L., Yip P.S.F. and Watson R.A.Y., A unified likelihood-based approach for estimating population size in continuous-time capture–recapture experiments with frailty, Biometrics. U.K., Blackwell Publishing, 2007, 63(1): 228-236.

Researcher : Xie S

List of Research Outputs

Yip P.S.F. and Xie S., Changes in marital structure and their impacts on Hong Kong society , Journal of Youth Studies. Hong Kong, The Hong Kong Federation of Youth Groups, 2006, 9(2): 14-29.

Researcher : Xu Y

List of Research Outputs

Yip P.S.F., Hsieh Y.H., Xu Y., Lam K.F., King C.C. and Chang H.L., Assessment of intervention measures for the 2003 SARS epidemic in Taiwan by use of a back-projection method, Infection Control and Hospital Epidemiology . University of Chicago Press, 2007, 28(5): 525-530.

Researcher : Yang CT

List of Research Outputs

Li C.K., Yang C.T., Poon N.L. and Fung T.W.K., Significance of sequence of strokes in Chinese handwriting examination, Journal of Forensic Sciences. USA, American Academy of Forensic Sciences, 2007, 52(2): 467-472.

Researcher : Yang H

Project Title:	Actuarial Science and Finance Models with Dependent Risks
Investigator(s):	Yang H
Department:	Statistics & Actuarial Science
Source(s) of Funding:	Competitive Earmarked Research Grants (CERG)
Start Date:	09/2005

Abstract:

The main objectives of this project are: (1) To study the asymptotic properties of the finite time horizon ruin probability, the distribution of the surplus before and after ruin, the joint distribution of the surplus immediately before and after ruin under our models and heavy tailed claim distributions. 2) Optimal Portfolio Selection under Correlated Asset Price Models. We will investigate the optimal asset allocation policy under correlated asset price models in both discrete and continuous settings. (3) Numerical Results. After obtaining some theoretical results, we would like to compare our results with those in the literature for the classical and various extended models.

Project Title:	Embedded options in insurance products and optimal policies for insurance portfolios under Markovian Regime switching models
Investigator(s):	Yang H, Yin GG
Department:	Statistics & Actuarial Science
Source(s) of Funding:	Competitive Earmarked Research Grants (CERG)
Start Date:	01/2007

Abstract:

(1) Pricing of embedded options in insurance products. (2) Optimal strategies for insurance portfolios. (3) Numerical and real data study.

List of Research Outputs

Gerber H.U., Lin X.S. and Yang H., A Note on the Dividends-penalty Identity and the Optimal Dividend Barrier, ASTIN Bulletin. Leuven, Peeters, 2006, 36: 489-503.

Lau J. .W., Siu T.K. and Yang H., On Bayesian Mixture Credibility, ASTIN Bulletin. Leuven, Peeters, 2006, 36: 573-588.

Li Z., Ng K.W., Tan K.S. and Yang H., Best CRP investment strategies for Dynamic Portfolio Selection, International Journal of Theoretical and Applied Finance. 2006, 9: 951-966.

Li Z.F., Yang H. and Deng X.T., Optimal Dynamic Portfolio Selection with Earnings-at-Risk, Journal of Optimization Theory and Applications. 2007, 132: 459-473.

Ng K.W., Yang H. and Zhang L., Upper Bounds for Ruin Probability in a filtered compound Poisson model , International Journal on Statistics and Systems. 2006, 1: 191-201.

Siu K.B.B. and Yang H., Expected shortfall under a model with market and credit risks, In: R. Mamon and R. Elliott, Hidden Markov Models in Finance. Springer, 2007, 91-100.

Yang H., 1st Runner-up of the ASHK Best Paper Award, The Actuarial Society of Hong Kong . 2006.

Yang H., Acta Mathematicae Applicatae Sinica. 2007.

Yang H., Encyclopedia of Quantitative Finance. 2007.

Yang H., IAENG International Journal of Applied Mathematics. 2007.

Yang H., Insurance; Mathematics and Economics. 2006.

Yang H., International Actuarial Meeting on Risk Measures and Solvency. Antalya, Turkey, 2007.

Yang H., International Journal on Statistics and Systems. 2006.

Yang H., Interpaly between actuarial science and finance, 2006 conference on financial mathematics and financial engineering. 2006.

Yang H., Optimal portfolio strategy under regime switching model, 10th International Congress on Insurance: Mathematics and Economics. Leuven, Belgium, 2006.

Yang H., Optimal portfolio strategy under regime switching model, Twelfth International Conferences on Computational and Applied Mathematics. Leuven, 2006.

Yang H., Recent Developments in Financial and Insurance Mathematics and the Interplay with the Industry. Oberwolfach, Germany, 2007.

Researcher : Yip PSF

Project Title:	Estimating population size for capture-recapture experiments with missing data and measurement error
Investigator(s):	Yip PSF
Department:	Statistics & Actuarial Science
Source(s) of Funding:	Competitive Earmarked Research Grants (CERG)
Start Date:	09/2005

Abstract:

The main objective of this project is to provide efficient estimation methods for estimating population size in capture-recapture experiments with incomplete information and possibly measurement error on covariates.

List of Research Outputs

Liu D.P., Yip P.S.F. and Huggins R.M., Nonparametric method for estimating the size of an open population using marginal data from repeated multiple lists, Australian & New Zealand Journal of Statistics . Australian Statistical Publishing Association Inc., 2007, 49(3): 303-320.

Ran M., Chen E.Y.H., Conwell Y., Chan C.L.W., Yip P.S.F., Xiang M. and Caine E.D., Mortality in people with schizophrenia in rural China: 10-year cohort study, British Journal of Psychiatry. 2007, 190: 237-242.

Xi L., Yip P.S.F. and Watson R.A.Y., A unified likelihood-based approach for estimating population size in continuous-time capture–recapture experiments with frailty, Biometrics. U.K., Blackwell Publishing, 2007, 63(1): 228-236.

Yip P.S.F., Hsieh Y.H., Xu Y., Lam K.F., King C.C. and Chang H.L., Assessment of intervention measures for the 2003 SARS epidemic in Taiwan by use of a back-projection method, Infection Control and Hospital Epidemiology . University of Chicago Press, 2007, 28(5): 525-530.

Yip P.S.F. and Xie S., Changes in marital structure and their impacts on Hong Kong society , Journal of Youth Studies. Hong Kong, The Hong Kong Federation of Youth Groups, 2006, 9(2): 14-29.

Yip P.S.F., Lee J. and Law C.K., Implications of the demographic transition, In: Gabreil M. Leung and John Bacon-Shone, Hong Kong's Health System: Reflections, Perspectives and Visions. Hong Kong, The Hong Kong University Press, HKU, 2006, 61-80.

Researcher : Yu PLH

Project Title:	Bayesian Analysis of Price Trend Models for Price Movements in Financial Markets
Investigator(s):	Yu PLH, Li WK
Department:	Statistics & Actuarial Science
Source(s) of Funding:	Seed Funding Programme for Basic Research
Start Date:	02/2005
Completion Date:	09/2006

Abstract:

Over the years, market investors and technical analysts have devised lots of technical indicators such as moving averages and momentum indicators to forecast the price trend of a financial market. Recent studies (for example van der Hart, et al. 2003, Osler, 2000 and Lo, et al. 2000) and reveal that these rules may provide positive net profits. This result clearly rejected the efficient market hypothesis, which claims that no additional benefit will be gained by using public information related to the market. In other words, asset prices cannot be forecasted from its historical prices or other past variables. The price trend models proposed by Taylor (1980) provide a good alternative solution and have produced many successful applications of the devised price trend trading rules in the futures markets and foreign exchange markets, see Taylor (1980, 1982), Taylor and Kingsman (1978), Taylor and Tari (1989) and Cabral and Guimaraes (1994). Since Taylor's foundation work on price trend model presented in 1980 was so useful for forecasting financial markets, his paper was reprinted as a book chapter in Mills (2002). Taylor (1980) proposed different versions of price trend model, the most basic form of which is called the basic price trend model as described below. Assume that the return R(t) at trading day t is the sum of two unobservable components: the trend T(t) and the white noise et, where the trend T(t) is equal to either the previous trend T(t-1) with probability p, or a new independent trend T(t) with probability 1-p. Therefore, the trend is a constant on each of a sequence of time intervals, each interval having a random duration. In addition, the quantity 1/(1-p) represents the expected trend duration and is a measure of the trend stability. This model assumes that the market satisfies the following conditions: (a) new information arrives randomly at the market; (b) there is new information on a proportion 1-p of the trading days; (c) trend changes only when new information becomes available; (d) when trend changes, the new trend is independent of all past trends; and (e) the trend values are determined by the current information about demand and supply. In estimating model parameters, Taylor applied the methods of moments based on the autocorrelation function and forecasts of future returns were generated as if the trend model is an ordinary ARMA(1,1) model (see Taylor 1986, 1992 and Cabral and Guimaraes 1994). However, such estimation and forecasting methods are rather ad hoc and do not have a strong theoretical basis. To improve the estimation and forecasting methods, Yu (PI) (with Kwan, Lam, So) (2000) developed the maximum likelihood method to estimate the parameters of the basic price trend model and derived one-step ahead forecast to predict the future trend. Based on these optimal forecasts, a new trading rule was constructed and was found to bear similarity to a popular trading rule based on moving averages. The empirical study on trading Hong Kong Hang Seng Index Futures indicated that the new trading rule could earn more net profit than the naïve buy-and-hold rule, especially in an up-and-down market. Time-varying volatility has already been identified as a well-known stylized fact in securities returns. It is thus worthwhile to extend Taylor's price trend models to incorporate time-varying volatility. This project aims at developing flexible price trend models to improve the trend prediction and the performance of the technical trading rules derived from the models. The first part of the proposed research is to develop price trend models in which the error variance is time-varying. The moments of the returns and stationary properties of the above models will be derived. We will adopt a Bayesian approach for parameter estimation of the proposed price trend models. We will also investigate how these models can be used to describe the microstructures of the financial markets. The second part of the proposed research will be devoted to the development of forecasting and trading strategies under the proposed models. We will compare the performances of the new trading rules with the commonly used technical rules.

Project Title:	Modeling of ranking data: a decision tree approach
Investigator(s):	Yu PLH
Department:	Statistics & Actuarial Science
Source(s) of Funding:	Competitive Earmarked Research Grants (CERG)
Start Date:	11/2005

Abstract:

To develop recursive partitioning algorithms for building binary tree models for ranking data; to develop hybrid decision tree models for ranking data by combining a decision tree model and a classical ranking model.

List of Research Outputs

Fung T.W.K. and Yu P.L.H., New-based learning in statistics, Proceedings of the 7th International Conference on Teaching Statistics. 2006.

Wu H., Yu P.L.H. and Li W.K., Value at risk estimation using independent component analysis-generalized autoregressive conditional heteroscedasticity (ICA-GARCH) models, International Journal of Neural Systems. Singapore, World Scientific, 2006, 16(5): 371-382.

Wu H.C. and Yu P.L.H., ICLUS: A robust and scalable clustering model for time series via independent component analysis, In: Peter Fleming, International Journal of Systems Science. UK, Taylor & Francis, 2006, 37(13): 987-1001.

Yu P.L.H., Integrating statistics and information technology into an investment risk course, Proceedings of the 7th International Conference on Teaching Statistics. 2006.

Researcher : Yuen KC

Project Title:	Analyses of insurance risk models with dividend payments
Investigator(s):	Yuen KC, Li WK, Ng KW
Department:	Statistics & Actuarial Science
Source(s) of Funding:	Competitive Earmarked Research Grants (CERG)
Start Date:	01/2006

Abstract:

To study several problems including the expection of discounted dividends until possible ruin and the ruin probability (possibly the classical risk model perturbed by diffusion); to investigate the optimal dividend strategy and then study the ruin probability, the deficit at ruin, and the surplus immediately before ruin (possibly the classical risk model with interest on negative surplus); to studty the Gerber-Shiu expected discounted penalty function which embraces some important actuarial quantities such as the ruin probability, the deficit at ruin, and the surplus immediately before ruin(possibly the classical risk model with investment income).

List of Research Outputs

Chen Y., Yuen K.C. and Ng K.W., Asymptotics for ruin probabilities of a two-dimensional renewal risk model with heavy-tailed claims , Research Report. Department of Statistics and Actuarial Science, HKU, 2007, 452: 1-12.

Shi J., Yuen K.C. and Xu J., Confidence bands for lifetime distribution functions with interval censored data , Research Report. Department of Statistics and Actuarial Science, HKU, 2006, 437: 1-14.

Wan L.M., Yuen K.C. and Li W.K., Analysis of an insurance risk model with thinning dependence and common shock, Journal of Actuarial Practice. U.S.A., Absalom Press, 2006, 13: 147-164.

Wang Y., He S., Zhu L. and Yuen K.C., Asymptotics for a censored generalized linear model with unknown link function, Probability Theory and Related Fields. 2007, 138(1): 235-267.

Yuen K.C., An introduction to some variants of the classical risk model, Year book, Statistical and Actuarial Science Society . HKU, 2006.

Yuen K.C., Wang G. and Wu R., On the renewal risk model with stochastic interest, Stochastic Processes and their Applications. 2006, 116(10): 1496-1510.

Yuen K.C., Ruin analysis of risk models with constant dividend barrier, 2007 Conference on Statistics, Finance and Actuarial Science, Kunming, China. 2007.

Yuen K.C., Some recent results for interval censored data, 2007 IASC-ARS Special Conference, Seoul, Korea. 2007.

Yuen K.C., Wang G. and Li W.K., The Gerber-Shiu expected discounted penalty function for risk processes with interest and a constant dividend barrier, Insurance: Mathematics and Economics. 2007, 40(1): 104-112.

Zhang Z., Li W.K. and Yuen K.C., On a mixture GARCH time series model, Journal of Time Series Analysis. Blackwell, 2006, 27(4): 577-597.

Researcher : Zhou J

List of Research Outputs

Hu Y., Zhou J. and Fung T.W.K., An extension of the transmission disequilibrium test incorporating imprinting, Genetics. USA, Genetics Society of America, 2007, 175: 1489-1504.

Hu Y., Zhou J., Sun F.Z. and Fung T.W.K., The transmission disequilibrium test and imprinting effects test based on case-parent pairs, Genetic Epidemiology. USA, Wiley-Liss, Inc., 2007, 31: 273-287.

Zhou J., Hu Y. and Fung T.W.K., A simple method for detection of imprinting effects based on case-parents trios, Heredity. UK, Nature Publishing Group, 2007, 98: 85-91.

Researcher : Zhou X

List of Research Outputs

Zhou X. and Li W.K., Modeling threshold conditional heteroscedasticity with regime-dependent skewness and kurtosis , Research Report. Department of Statistics and Actuarial Science, HKU, 2007, 448: 1-28.

-- End of Listing --