SS1 | Spectral analysis of time series: classical methods, wavelets and soft-computing methodologies The spectral analysis of time series is of paramount importance in many scientific disciplines: geosciences, physics, medicine, astronomy, economics, ... to name just a few. The first methods were based on Fourier series but were extended later to parametric approaches like maximum entropy. More recently, wavelets have become very popular because they have expanded the range of signals than can be analysed. Even more recently, soft computing techniques, like neural networks computing, have allowed to relax the assumptions required by many methods. This session covers both theoretical developments and practical applications of the spectral analysis of time series. Classical methods like Fourier methods, Blackman-Tuckey, maximum entropy, multi-taper methods as well as wavelets and soft-computing methodologies are considered. Particular problems of spectral analysis like time series with uneven sampling, compositional data time series, time series of averaged or integrated values and non-Gaussian time series are of special interest. Organizers: Eulogio Pardo-Igúzquiza, Instituto Geológico y Minero de España (IGME), Madrid (Spain). Francisco Javier Rodríguez-Tovar, Facultad de Ciencias, Universidad de Granada, Granada (Spain). |
---|
SS2 | Inference via Estimating Functions for circular and count time series models Recently there is interest in using estimating functions for making inference for econometrics and financial modeling with option pricing applications. Thavaneswaran and Abraham (1988) introduced a unified theory of estimating functions to study inference for linear and non linear time series models with finite variance and demonstrated the superiority of the approach. Since then many papers have been written, illustrating the use of estimating functions for different types of time series models and stochastic volatility models. There has also been growing interest in stochastic processes with infinite variance and circular time series. This is due to the inherent challenge and theoretical interest provided by the non-normal stable laws as well as the possibility that the processes constructed from these laws may be appropriate models for many diverse phenomena. In practice, any time series which exhibits sharp spikes or occasional bursts of outlying observations suggests the possible use of a model with stable errors having infinite variance. For time series models with infinite-variance stable errors and for circular time series time series with wrapped stable errors, closed form expressions for the density are not available and hence the maximum likelihood estimate cannot be obtained. Merkouris (2007) and Thavaneswaran et al. (2013) have used combined sine and cosine estimating functions to study estimation and recursive estimation. Circular time series and integer valued time series have become topics of great interest recently. In this special invited session, we propose to discuss inference for circular time series models, count time series models and the use of estimating functions for inference in such models. Speakers: Aera Thavaneswaran, Dept. of Statistics, University of Manitoba, Winnepeg, Manitoba, Canada. Title: Estimating functions for circular time series. Ibrahim Bin Mohamed, University of Malaysia, Kualalumpur, Malaysia. Title: Outliers in a circular time series. Melody Ghahramani, Department of Mathematics & Statistics, University of Winnepeg, Winnipeg, Manitoba, Canada. Title: Semi-parametric estimation of count time series. Organizers: Bovas Abraham, University of Waterloo, Waterloo, Ontario, Canada. |
---|
Special Talk | George Box: The 'Accidental Statistician' who revolutionized time series analysis George Edward Pelham Box was born on October 19, 1919 in Gravesend, Kent, England and died on March 28, 2013 in Madison, Wisconsin. George Box was one of the world's most leading statisticians and made path breaking contributions to many areas of statistics including design of experiments, robustness, Bayesian methods, time series analysis and forecasting, and quality improvement. This talk discusses his contributions to time series analysis and forecasting. His work in this area started in collaboration with Gwilym Jenkins in the early 1960's and continued over the next several decades. His contributions include the classic and extraordinarily influential book "Time Series Analysis: Forecasting and Control" written with Gwilym Jenkins and first published by Holden Day in 1970. His subsequent contributions to time series analysis include joint work with George Tiao and many former graduate students including this one. His work provided a unified framework for carrying out time series analysis in practice and laid the foundation for many new developments in the field. In this discourse, I will discuss many significant developments in time series analysis including ARIMA models and the time series modeling strategy proposed by Box and Jenkins (1970), along with a discussion of their contributions to forecasting. We will consider the use of these models for assessing the impact of external interventions as described by Box and Tiao (1975). Extensions of this approach to detect outliers will also be considered (for example, Abraham and Box (1979)). Box and Newbold (1971) commented on spurious correlations in the context of correlating non-stationary time series, and Box and Tiao (1977) discussed canonical analysis of multiple time series. The ideas in these two papers laid the foundation for the work on co-integration by Rob Engle and Clive Granger who won the Nobel Prize in Economics in 2003. George Box's work has had a great impact on the theory and practice of time series analysis. Organizer and Speaker: Bovas Abraham, University of Waterloo, Waterloo, Ontario, Canada. |
---|