Modeling the Error Term by Moving Average and Generalized Autoregressive Conditional Heteroscedasticity Processes | Science Publications

American Journal of Applied Sciences

Ayodele Abraham Agboluaje¹  Chee Yin Yip¹  Suzilah bt Ismail¹ 

关键词: Determinant Residual Covariance;    Minimum Forecast Errors;    Minimum Information Criteria;    Leverage;    Log Likelihood;   

DOI  :  10.3844/ajassp.2015.896.901

学科分类：自然科学（综合）

来源: Science Publications

【 摘 要 】

This study has been able to reveal that the Combine WhiteNoise model outperforms the existing Generalized Autoregressive ConditionalHeteroscedasticity (GARCH) and Moving Average (MA) models in modeling theerrors, that exhibits conditional heteroscedasticity and leverage effect. MAprocess cannot model the data that reveals conditional heteroscedasticity andGARCH cannot model the leverage effect also. The standardized residuals of GARCH errors are decomposed into series ofwhite noise, modeled to be Combine White Noise model (CWN). CWN modelestimation yields best results with minimum information criteria and high loglikelihood values. While the EGARCH model estimation yields betterresults of minimum information criteria and high log likelihood values whencompare with MA model. CWN has the minimum forecast errors which areindications of best results when compare with the GARCH and MA models dynamicevaluation forecast errors. Every result of CWN outperforms the results of bothGARCH and MA.

【 授权许可】

Unknown