Institute for Mathematical Sciences Event Archive
Online registration form Enquiries |
|
Workshop on Recent Advances in Nonlinear Time Series Analysis
(7 - 18 February 2011)
Organizing Committee · Visitors and Participants · Overview · Activities · Venue
Chair
- Howell Tong (London School of Economics)
Co-Chairs
- Ying Chen (National University of Singapore)
- Yingcun Xia (National University of Singapore)
In Statistics, Nonlinear Time Series Analysis started around 1980. After a slow start, its development has been rapid recently. The following events have provided timely and important momentum to the development of the subject.
- Research Workshop on Nonlinear Time Series Analysis and Applications Held in Edinburgh on 12-25 July 1989 (with 10 papers published in a special issue on Nonlinear Time Series Analysis, Statistica Sinica, volume 1, 1991);
- The Royal Statistical Society Chaos Day in London in 1991 (with 7 papers and discussions published in Series B of the Journal of the RSS in 1992);
- The Royal Society (London) Discussion Meeting on Chaos and Forecasting in 1994 (with 15 papers and discussions published in the Transactions of the Royal Society, 1994);
- The International Workshop on Financial Statistics in Hong Kong in 1999 (with proceedings published by Imperial College Press/World Scientific in 2000);
- The International conference on Threshold Models and New Developments in Time Series, celebrating Howell Tong’s 60th birthday, in Hong Kong in 2004 (with 13 papers published in Statistica Sinica, 2007).
Important specialist monographs have also appeared steadily over the past twenty years or so, notably Priestley (1989), Tong (1983, 1990), Chan and Tong (2001), Fan and Yao (2003) and others.
The need to break away from linear models was recognized almost as early as 1927 when Udny Yule introduced the first linear time series model, namely the autoregressive model. The fact that it took statisticians more than half a century to accomplish the first breakthrough might seem strange at first sight. However, if we bear in mind the following historical perspectives, then we might begin to appreciate the enormity of the task. The inevitable complexity of a nonlinear time series model necessitates computing power not easily available before the 1970s. The relevant theory in probability theory to handle the ergodicity/stationarity issues of a Markov chain over Rp was not available till around the mid-1970s. Some relevant statistical methodologies, such as model selection, conditional least squares and others, were only fully developed in the 1970s. In the deterministic world, even the notion of chaos had to wait till 1976 before it took off.
After the initial breakthroughs in the early 1980s, progress was still slow over the decade, due to the enormity of the theoretical and methodological challenges. Fortunately, with successful applications in diverse fields, e.g. ecology, economics, epidemiology, finance, hydrology and many others, and interaction with dynamicists, an increasing number of bright and young statisticians started to be attracted to the exciting area in the late 1980s and early 1990s. As a result, publications have grown exponentially. Just as an illustration, according to the ISI Web of Knowledge, the citations of a threshold time series model have reached about 1,000 in 2009 since its birth in 1980. The corresponding figures in 1990 and 2000 were approximately 10 and 250 respectively!
Many exciting developments are indeed taking place. The relatively recent entry of non-parametric and semi-parametric techniques into the time series analysis has given modern time series analysts powerful tools to explore the time series data from many different perspectives, before deciding on an appropriate parametric model for more in-depth exploration. Nonlinear dimension reduction methodology is being developed and exploited to handle high-dimensional time series data. Various factor models are being developed to cope with high frequency data, so often found in financial applications. The tantalizing connection between nonlinearity and long-memory is yet to be fully investigated. The connection between non-linear stochastic differential equations and discrete-time nonlinear time series models awaits careful study. The almost unique signature of nonlinear time series analysis is its challenge to our ability to bring the most relevant probability theory, statistical methodology, dynamical system knowledge and computing ingenuity to bear so as to get a clearer glimpse at the dynamical world around us.
In Singapore and especially at the NUS, there are clusters of expertise in the methodology of nonlinear time series analysis with particular reference to applications in finance, infectious diseases and environmental impacts on public health. Some of the local expertise has reached the international level and others approaching it.
The main mission of the workshop is to bring together international researchers with expertise in the different areas listed earlier, namely non-parametrics, semi-parametrics, dimension reduction, high-dimensional time series, and many others, so as to gain a deeper and wider understanding of the dynamical world. By participating in this exciting venture, local nonlinear enthusiasts will have ample opportunities of interacting with international experts at the highest level. There is no doubt that the workshop will be an important milestone in the development of nonlinear time series analysis on an international scale. It will also help bring about a quantum leap in research in non-linear time series analysis in Singapore.
Tuesday, 8 Feb 2011 |
|
09:15am - 09:30am |
Registration |
09:30am - 09:35am |
Opening Remarks |
|
Session I |
09:35am - 12:30pm |
Feature matching in time series modelling Howell Tong, London School of Economics, UK Estimating extremal dependence in time series via the extremogram Richard Davis, Columbia University, USA |
12:30pm - 02:00pm |
--- Lunch --- |
|
Session II |
02:00pm ~ |
Modeling and estimation for nonstationary financial time series Ying Chen,National University of Singapore |
Wednesday, 9 Feb 2011 |
|
09:50am - 10:00am |
Registration |
|
Session I |
10:00am - 12:30pm
|
Particle filter based on-line estimation of spot and cross volatility with non-linear market micro-structure noise models (PDF) Rainer Dahlhaus, University of Heidelberg, Germany
Stochastic covariance models (PDF) Mike, Ka Pui So, Hong Kong University of Science and Technology, Hong Kong |
12:30pm - 02:00pm |
--- Lunch --- |
|
Session II |
02:00pm ~ |
Inference for some continuous-time models in finance Peter Brockwell, Colorado State University, USA and University of Melbourne, Australia |
Thursday, 10 Feb 2011 |
09:50am - 10:00am |
Registration |
|
Session I |
10:00am - 12:30pm |
A new fractional IGARCH model Guodong Li, University of Hong Kong, Hong Kong Quantile estimation of threshold autoregressive models with exogenous variables and heteroskedasticity Cathy W. S. Chen, Feng Chia University, Taiwan |
12:30pm - 02:00pm |
--- Lunch --- |
|
Session II |
02:00pm ~ |
Plague Noelle Samia, Northwestern University, USA |
Friday, 11 Feb 2011 |
|
09:50am - 10:00am |
Registration |
|
Session I |
10:00am - 12:30pm |
Statistical modelling of nonlinear long-term cumulative effects Yingcun Xia, National University of Singapore
Factor modelling for high dimensional time series Qiwei Yao, London School of Economics, UK |
12:30pm - 02:00pm |
--- Lunch --- |
|
Session II |
02:00pm ~ |
A general framework for identification of time-varying dynamic systems using multiple models Cheng Xiang, National University of Singapore |
Monday, 14 Feb 2011 |
|
09:50am - 10:00am |
Registration |
|
Session I |
10:00am - 11:30am |
Modelling nonlinear time series with spatial neighbouring effects: a personal review Zudi Lu, University of Adelaide, Australia |
11:30am - 02:00pm |
--- Welcome lunch reception at IMS --- |
|
Session II |
02:00pm - 03:30pm |
Goodness of fit tests for time series based on scores (PDF) Shiqing Ling, Hong Kong University of Science and Technology, Hong Kong |
Tuesday, 15 Feb 2011 |
|
09:50am - 10:00am |
Registration |
|
Session I |
10:00am - 11:30am |
On least squares estimation of multiple-regime TAR models (PDF) Dong Li, Hong Kong University of Science and Technology, Hong Kong |
11:30am - 02:00pm |
--- Lunch --- |
|
Session II |
02:00pm - 03:30pm |
A backward stochastic differential equation approach to convex risk measures for derivative securities Ken Siu, Macquarie University, Australia |
Wednesday, 16 Feb 2011 |
|
09:50am - 10:00am |
Registration |
|
Session I |
10:00am - 11:30am |
Estimation of a threshold autoregressive model under misspecification Myung Hwan Seo, London School of Economics, UK |
11:30am - 02:00pm |
--- Lunch --- |
|
Session II |
02:00pm - 03:30pm |
Quasi-likelihood estimation of threshold diffusion processes Kung-Sik Chan, University of Iowa, USA |
Thursday, 17 Feb 2011 |
|
09:50am - 10:00am |
Registration |
|
Session I |
10:00am - 11:30am |
Quantiles, spectral analysis and time series Marc Hallin, Université Libre de Bruxelles, Belgium |
11:30am - 02:00pm |
--- Lunch --- |
|
Session II |
02:00pm - 03:30pm |
Mixture Kalman filter and plug-and-play navigation systems Rong Chen, Rutgers University, USA |
Friday, 18 Feb 2011 |
|
09:50am - 10:00am |
Registration |
|
Session I |
10:00am - 11:00am |
Test for nonlinearity based on entropy measure Simone Giannerini, University of Bologna, Italy |
11:00pm - 11:30am |
--- Coffee Break --- |
11:30am - 12:30pm |
Trending time seies and non- and semi-parametric cointegration Jiti Gao, The University of Adelaide, Australia |
12:30pm - 02:00pm |
--- Lunch --- |
|
Open Forum |
02:00pm - 03:30pm |
Open Forum I: Future directions of nonlinear time series analysis
Open Forum II: More future directions of nonlinear time series analysis |
Students and researchers who are interested in attending these activities and who do not require financial aid are requested to complete the online registration form.
The following do not need to register:
- Those invited to participate.
- Those applying for financial support.
Organizing Committee · Visitors and Participants · Overview · Activities · Venue
