Activity Number:
|
580
- Time Series and Factor Models
|
Type:
|
Contributed
|
Date/Time:
|
Wednesday, August 2, 2017 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Business and Economic Statistics Section
|
Abstract #322532
|
View Presentation
|
Title:
|
In-Curve Updating of Predictions for Functional Time Series
|
Author(s):
|
Shuhao Jiao* and Alexander Aue
|
Companies:
|
Department of Statistics, UC Davis and University of California, Davis
|
Keywords:
|
Functional data ;
Prediction ;
Principal components ;
Time series
|
Abstract:
|
Functional time series have become an important tool when dealing with a data that may conveniently be viewed as a curves exhibiting a temporal dependence structure. Such examples are found in variety of application areas. Here we focus on an environmental data set on daily pollution concentration curves. It is discussed how existing functional prediction methodology can be updated in an automatic way using in-curve updates that improve predictions obtained from time series and linear model formulations of the problem. Simulation results and real data examples highlight the performance of the proposed procedure.
|
Authors who are presenting talks have a * after their name.