|
Activity Number:
|
477
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Business and Economic Statistics Section
|
| Abstract - #304271 |
|
Title:
|
Application of Variable Selection Method via Spectral Analysis to Seasonal Linear Models in Some Call Volume Forecasting
|
|
Author(s):
|
Myung Suk Kim*+ and Taek Soo Shin
|
|
Companies:
|
Sogang University and Citi Bank
|
|
Address:
|
#1, Shinsu-dong, Mapo-gu, Seoul, 121-742, Korea
|
|
Keywords:
|
Call Volume Forecast ; Seasonal Linear Model ; Seasonal Variable Selection ; Spectral Analysis
|
|
Abstract:
|
Many call volume forecasting models include various seasonal variables. However, a rigorous variable selection method has not seem to be applied for the seasonal variable selection. Furthermore, the importance of interaction effect by several seasonal variables has been underestimated. In this paper a seasonal variable selection method via the spectral analysis is applied to seasonal linear models with some interaction effect. Our empirical results indicate that updated seasonal linear models with newly selected variables seem to outperform existing seasonal linear models in one day ahead and one week ahead call volume forecasts under various prediction criteria.
|
