A0283
Title: Interval estimation for continuous-time correlation
Authors: Philip Reiss - University of Haifa (Israel) [presenting]
Biplab Paul - University of Haifa (Israel)
Noemi Foa - University of Haifa (Israel)
Dror Arbiv - University of Haifa (Israel)
Abstract: Continuous-time correlation is a recently proposed way to measure association between time series that are noisy and may be irregularly observed. The crux of the method is basis-function smoothing of the time series, which can effectively mitigate the well-known attenuation problem for correlation estimation with noisy data. This technique is reminiscent of functional data analysis, but treats the observations, rather than the variables, as forming a continuum. This talk will focus on interval estimation for continuous-time correlation. We present a bootstrapping method and several variants of posterior simulation, which entail different assumptions about between-curve error dependence. Some of these approaches allow the variables to be observed at different time points. Moving beyond inferring the correlation between two curves observed with noise, we also consider inference for the correlation parameter of the underlying bivariate stochastic process. The methods are validated by simulation, and are illustrated with incompletely observed international development data and with electroencephalography data.
