Application of an EWMA combining technique to the prediction of currency exchange rates

Hyung Won Shin, So Young Sohn

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

Financial forecasting is an important and challenging task for both academic researchers and business practitioners. The recent trend to improve the prediction accuracy is to combine individual forecasts using a simple average or weighted average where the weight reflects the inverse of the prediction error. In the existing combining methods, however, the errors between actual and predicted values are equally reflected in the weights regardless of the time order in a forecasting horizon. In this paper, we propose a new approach where the forecasting results of Generalized AutoRegressive Conditional Heteroskedastic (GARCH), neural network, and random walk models are combined based on a weight that reflects the inverse of the exponentially weighted moving average of the Mean Absolute Percentage Error (MAPE) of each individual prediction model. The results of an empirical study indicate that the proposed method has a better accuracy than the GARCH, neural network, and random walk models, and also combining methods based on using the MAPE for the weight.

Original languageEnglish
Pages (from-to)639-644
Number of pages6
JournalIIE Transactions (Institute of Industrial Engineers)
Volume39
Issue number6
DOIs
Publication statusPublished - 2007 Jun 1

Fingerprint

Neural networks
Industry

All Science Journal Classification (ASJC) codes

  • Industrial and Manufacturing Engineering

Cite this

@article{9aff2c8504774e1ebe8b3b96b8b31738,
title = "Application of an EWMA combining technique to the prediction of currency exchange rates",
abstract = "Financial forecasting is an important and challenging task for both academic researchers and business practitioners. The recent trend to improve the prediction accuracy is to combine individual forecasts using a simple average or weighted average where the weight reflects the inverse of the prediction error. In the existing combining methods, however, the errors between actual and predicted values are equally reflected in the weights regardless of the time order in a forecasting horizon. In this paper, we propose a new approach where the forecasting results of Generalized AutoRegressive Conditional Heteroskedastic (GARCH), neural network, and random walk models are combined based on a weight that reflects the inverse of the exponentially weighted moving average of the Mean Absolute Percentage Error (MAPE) of each individual prediction model. The results of an empirical study indicate that the proposed method has a better accuracy than the GARCH, neural network, and random walk models, and also combining methods based on using the MAPE for the weight.",
author = "Shin, {Hyung Won} and Sohn, {So Young}",
year = "2007",
month = "6",
day = "1",
doi = "10.1080/07408170600899474",
language = "English",
volume = "39",
pages = "639--644",
journal = "IISE Transactions",
issn = "2472-5854",
publisher = "Taylor and Francis Ltd.",
number = "6",

}

Application of an EWMA combining technique to the prediction of currency exchange rates. / Shin, Hyung Won; Sohn, So Young.

In: IIE Transactions (Institute of Industrial Engineers), Vol. 39, No. 6, 01.06.2007, p. 639-644.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Application of an EWMA combining technique to the prediction of currency exchange rates

AU - Shin, Hyung Won

AU - Sohn, So Young

PY - 2007/6/1

Y1 - 2007/6/1

N2 - Financial forecasting is an important and challenging task for both academic researchers and business practitioners. The recent trend to improve the prediction accuracy is to combine individual forecasts using a simple average or weighted average where the weight reflects the inverse of the prediction error. In the existing combining methods, however, the errors between actual and predicted values are equally reflected in the weights regardless of the time order in a forecasting horizon. In this paper, we propose a new approach where the forecasting results of Generalized AutoRegressive Conditional Heteroskedastic (GARCH), neural network, and random walk models are combined based on a weight that reflects the inverse of the exponentially weighted moving average of the Mean Absolute Percentage Error (MAPE) of each individual prediction model. The results of an empirical study indicate that the proposed method has a better accuracy than the GARCH, neural network, and random walk models, and also combining methods based on using the MAPE for the weight.

AB - Financial forecasting is an important and challenging task for both academic researchers and business practitioners. The recent trend to improve the prediction accuracy is to combine individual forecasts using a simple average or weighted average where the weight reflects the inverse of the prediction error. In the existing combining methods, however, the errors between actual and predicted values are equally reflected in the weights regardless of the time order in a forecasting horizon. In this paper, we propose a new approach where the forecasting results of Generalized AutoRegressive Conditional Heteroskedastic (GARCH), neural network, and random walk models are combined based on a weight that reflects the inverse of the exponentially weighted moving average of the Mean Absolute Percentage Error (MAPE) of each individual prediction model. The results of an empirical study indicate that the proposed method has a better accuracy than the GARCH, neural network, and random walk models, and also combining methods based on using the MAPE for the weight.

UR - http://www.scopus.com/inward/record.url?scp=33947657340&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33947657340&partnerID=8YFLogxK

U2 - 10.1080/07408170600899474

DO - 10.1080/07408170600899474

M3 - Article

AN - SCOPUS:33947657340

VL - 39

SP - 639

EP - 644

JO - IISE Transactions

JF - IISE Transactions

SN - 2472-5854

IS - 6

ER -