B0899
Title: A historical functional linear model in the spatiotemporal setting of a flowing river
Authors: Craig Wilkie - University of Glasgow (United Kingdom) [presenting]
Surajit Ray - University of Glasgow (United Kingdom)
Claire Miller - University of Glasgow (United Kingdom)
Marian Scott - University of Glasgow (United Kingdom)
Abstract: For a regression setting where observations are indexed by time, the historical functional linear model accounts for the directionality of time, where the response at time t is affected by values of the covariate before t. In rivers, the directionality of water flow suggests that distance along the river from the source could be treated similarly to time. The historical functional linear model is fitted that accounts for both the directionality of time and flow to satellite remote sensing chlorophyll data for the Ramganga river in India. Results are presented to compare its performance to other modelling frameworks that do not explicitly account for this directionality.
