The Frank Hansford-Miller Fellow in 2023 is Prof. Guy Nason of Imperial College London.
Network time Series
Professor Guy Nason, Imperial College London.
A network time series is a multivariate time series where the individual series are known to be linked by some underlying network structure. Sometimes this network is real and known beforehand, but sometimes one assumes an unknown virtual network and tries to infer it along with the modelling.
Network time series are becoming increasingly common, long and collected over a massive number of variables. There are many variants including network time series whose underlying network changes over time, multiple network time series, series with different link types, antagonistic network time series, count data series to name but a few.
This talk advertises some recent methods for the modelling and fitting of network time series, pros and cons and, in some circumstances, attractive forecasting performance. We plan to show examples from epidemiology and economics, sometimes both together.
About the Speaker
Guy Nason is Chair in Statistics at Imperial College London. He has undertaken research in the areas of wavelets and time series analysis, particularly in the area of non-stationary time series.
A major current research direction is network time series in which he is involved in the NeSTProgramme Grant - a partnership of nine investigators at the six UK universities of Bath, Bristol, LSE, Oxford, York and Imperial. He is particularly interested in the statistical modelling of wine production and is especially interested in modelling the wines of Western Australia.