A0669
Title: Long memory network time series
Authors: Chiara Boetti - University of Bath (United Kingdom) [presenting]
Matthew Nunes - University of Bath (United Kingdom)
Marina Knight - University of York (United Kingdom)
Abstract: Many scientific areas, from computer science to the environmental sciences and finance, give rise to multivariate time series that exhibit long memory. Efficient modeling and estimation in such settings is key for a number of analysis tasks, such as accurate prediction. However, traditional approaches for modeling such data, for example, long memory vector autoregressive processes, are challenging even in modest dimensions, as the number of parameters grows quadratically with the number of modeled variables. In many data settings, the observed series is accompanied by a (possibly inferred) network that provides information about the presence or absence of between-component associations. Two new models are proposed for capturing the dynamics of long memory time series, where a network is taken into consideration. The approach not only facilitates the analysis of graph-structured long memory time series, but also improves computational efficiency over traditional multivariate long memory models by leveraging the inherent low-dimensional parameter space. Likelihood-based estimation algorithms are adapted to the network setting. Simulation studies show that the parameter estimation is more stable than traditional models and is able to tackle data scenarios where these models fail due to computational challenges. The efficacy of the proposed models is demonstrated on datasets arising in environmental science and finance applications.
