BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Regression with Dependent Functional Errors-in-Predictors - Xingha
 o Qiao (London School of Economics)
DTSTART:20180517T100000Z
DTEND:20180517T110000Z
UID:TALK107173@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:Functional regression is an important topic in functional data
  analysis. Traditionally\, in functional regression\, one often assumes th
 at samples of the functional predictor are independent realizations of an 
 underlying stochastic process\, and are observed over a grid of points con
 taminated by independent and identically distributed measurement errors. H
 owever\, in practice\, the dynamic dependence across different curves may 
 exist and the parametric assumption on the measurement error covariance st
 ructure could be unrealistic. In this paper\, we consider functional linea
 r regression with serially dependent functional predictors\, when the cont
 amination of predictors by measurement error is "genuinely functional" wit
 h fully nonparametric covariance structure. Inspired by the fact that the 
 autocovariance operator of the observed functional predictor automatically
  filters out the impact from the unobservable measurement error\, we propo
 se a novel generalized-method-of-moments estimator of the slope function. 
 The asymptotic properties of the resulting estimators under different scen
 arios are established. We also demonstrate that the proposed method signif
 icantly outperforms possible competitors through intensive simulation stud
 ies. Finally\, the proposed method is applied to a public financial datase
 t\, revealing some interesting findings. This is a joint work with Cheng C
 hen and Shaojun Guo.
LOCATION:Seminar Room 2\, Newton Institute
END:VEVENT
END:VCALENDAR