Article: Transformative advances in volatility prediction: unveiling an innovative model selection method using exponentially weighted information criteria Journal: International Journal of Business and Systems Research (IJBSR) 2024 Vol.18 No.6 pp.569 - 590 Abstract: Using information criteria is a common method for making a decision about which model to use for forecasting. There are many different methods for evaluating forecasting models, such as MAE, RMSE, MAPE, and Theil-U, among others. After the creation of AIC, AICc, HQ, BIC, and BICc, the two criteria that have become the most popular and commonly utilised are Bayesian IC and Akaike's IC. In this investigation, we are innovative in our use of exponential weighting to get the log-likelihood of the information criteria for model selection, which means that we propose assigning greater weight to more recent data in order to reflect their increased precision. All research data is from the major stock markets' daily observations, which include the USA (GSPC, DJI), Europe (FTSE 100, AEX, and FCHI), and Asia (Nikkei). Inderscience Publishers - linking academia, business and industry through research

Title: Transformative advances in volatility prediction: unveiling an innovative model selection method using exponentially weighted information criteria

Authors: Youyuan Wu; Wei Chong Choo; Bolaji Tunde Matemilola; Wan Cheong Kin; Zhe Zhang

Addresses: School of Business and Economics, University Putra Malaysia, Serdang, 43400, Malaysia ' School of Business and Economics, University Putra Malaysia, Serdang, 43400, Malaysia; Laboratory of Computational Statistics and Operations Research, Institute for Mathematical Research, University Putra Malaysia, Serdang, 43400, Malaysia ' Faculty of Economic and Management, University Putra Malaysia, Serdang, 43400, Malaysia ' Tunku Abdul Rahman University of Management and Technology, Malaysia ' School of Business and Economics, University Putra Malaysia, Serdang, 43400, Malaysia

Abstract: Using information criteria is a common method for making a decision about which model to use for forecasting. There are many different methods for evaluating forecasting models, such as MAE, RMSE, MAPE, and Theil-U, among others. After the creation of AIC, AICc, HQ, BIC, and BICc, the two criteria that have become the most popular and commonly utilised are Bayesian IC and Akaike's IC. In this investigation, we are innovative in our use of exponential weighting to get the log-likelihood of the information criteria for model selection, which means that we propose assigning greater weight to more recent data in order to reflect their increased precision. All research data is from the major stock markets' daily observations, which include the USA (GSPC, DJI), Europe (FTSE 100, AEX, and FCHI), and Asia (Nikkei).

Keywords: exponential weighted information criteria; volatility forecasting; decision making.

DOI: 10.1504/IJBSR.2024.142085

International Journal of Business and Systems Research, 2024 Vol.18 No.6, pp.569 - 590

Received: 19 Feb 2024
Accepted: 29 Mar 2024
Published online: 07 Oct 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Transformative advances in volatility prediction: unveiling an innovative model selection method using exponentially weighted information criteria

Keep up-to-date