A0281
Title: Non-parametric inference of spatial birth-death-move processes
Authors: Ronan Le Guevel - University of Rennes (France) [presenting]
Frederic Lavancier - University of Nantes (France)
Abstract: The analysis of the spatiotemporal dynamics of proteins involved in exocytosis in cells is an important biological challenge. A new spatial birth-death-move process, where the birth and death dynamics depend on the current spatial configuration of the population and where individuals can move with possible interactions during their lifetime, is introduced in order to model this complex mechanism. A non-parametric kernel method, with continuous time observations as well as discrete-time observations, is presented for the estimation of the jump intensity function, involving, in most cases, a choice of a distance between point configurations and a choice of a window which will also be discussed. The procedure is illustrated with simulations and results on the biological dataset.
