| Research Projects Supported by HKU's High Performance Computing Facilities |
| Researcher: |
| Dr Wai-sum Chan, Department of Statistics and Actuarial Science |
| Project Title: |
| Some Nonlinear Time-Series Models for Actuarial Use |
| Project Description: |
|
In this research project, we propose two advanced nonlinear time-series modelling techniques that actuaries can use in pricing insurance products, analysing results from experience studies, and forecasting actuarial assumptions. Nonlinear time-series models are able to provide much more flexibility as compared to classical linear ARIMA models. Unlike Bayesian methods, our proposed approaches do not require subjective specification of prior distributions. |
| Project Duration: |
| 36 months |
| Project Significance: |
|
The major contributions of this research project to actuarial research and
education are: • to introduce actuaries some advanced nonlinear time-series techniques that might be useful in building stochastic models for pricing and reserving; and • to illustrate these techniques step-by-step so that actuaries or actuarial students not expert in this area can still perform the procedures. |
| Results Achieved: |
|
• The project received funding from the U.S. Actuarial Education & Research
Fund (AERF): US$7,000. • Two M.Phil. students trained (Miss Kam Po Ling, 2003; and Miss Ng Man Wai, 2001). • Two manuscripts written: 1. On Robustness of Threshold-type of Non-linearity Tests, Journal of Forecasting, accepted for publication. 2. Mixture Time-Series Models with Heavy Tailed Distributions. |
| Remarks on the Use of High Performance Computing Cluster: |
| This research project involves large-scale simulations and computer-intensive E.M. Algorithms for estimation. The project would not be viable without the HPC Cluster. |
| Email Address: |
| chanws@hku.hk |