Abstract Details
Activity Number:
|
575
|
Type:
|
Contributed
|
Date/Time:
|
Wednesday, August 6, 2014 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Section on Statistics and the Environment
|
Abstract #311947
|
|
Title:
|
Forecasting Nonstationary Energy Time Series
|
Author(s):
|
Marina Knight*+ and Rebecca Killick and Guy Nason and Idris Eckley
|
Companies:
|
University of York and Lancaster University and University of Bristol and Lancaster University
|
Keywords:
|
forecasting ;
nonstationarity ;
wavelets
|
Abstract:
|
Within the energy sector forecasting is an important statistical tool. Each day many forecasts are made across a variety of time scales, such as production of renewables, consumer demand and trader pricing. Traditional statistical techniques assume stationarity of the past in order to produce accurate forecasts. For data arising in the energy sector, this stationarity assumption is often violated- hence certain conditions need to be imposed on the time-varying structure of the process in order to achieve meaningful estimation. As many real-life data display characteristics from the locally stationary (LS) framework, we propose a new forecasting method for LS time series. Our approach hinges on introducing a new measure to characterise the behaviour of a LS wavelet process: the localised partial autocorrelation function. While the partial autocorrelation function is widely used in the stationary world, for nonstationary processes this is the first time when such a measure is defined, investigated and exploited. We discuss the theoretical properties of our estimators, illustrate our method with a simulation study and show improved forecasting errors using this new technique.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2014 program
|
2014 JSM Online Program Home
For information, contact jsm@amstat.org or phone (888) 231-3473.
If you have questions about the Professional Development program, please contact the Education Department.
The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.
Copyright © American Statistical Association.