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 #323923
|
View Presentation
|
Title:
|
Dynamic Factor Model for Functional Time Series
|
Author(s):
|
Israel Martínez Hernández* and Jesús Gonzalo and Graciela González Farías
|
Companies:
|
and Universidad Carlos III de Madrid and Centro de Investigación en Matemáticas, CIMAT
|
Keywords:
|
Functional Time Series ;
Dynamic Factor Model ;
Non-Stationarity
|
Abstract:
|
In econometrics, identify the factors that drives the behavior of observed time series is important, even more when the data is functional. In this study we propose a new approach the Dynamic Factor Model for Functional Time Series. Our model will characterize the dynamic of the data through the factors and the functional behavior of the data via the factor loadings. Via simulation studies, we show that the factors capture the dynamic of the process and share the same properties of the functional data, such as stationarity. We propose a methodology to estimate the factors that produces a consistent estimator and we investigate finite sample performance using Monte Carlo studies.
|
Authors who are presenting talks have a * after their name.