Research has substantiated that the presence of outliers in data usually introduces additional errors and biases, which typically leads to a degradation in the precision of volatility forecasts. However, correcting outliers can mitigate these adverse effects. This study corrects the additive outliers through a weighting method and let these corrected values to replace the original outliers. Then, the model parameters are re-estimated based on this new return series. This approach reduces the extent to which outliers distort volatility estimates, allowing the model to better adapt to market conditions and improving the accuracy of volatility forecasts. This study introduces this approach for the first time to generalized autoregressive conditional heteroskedasticity mixed data sampling (GARCH-MIDAS) models, so as to establish an additional outliers corrected GARCH-MIDAS model (AO-GARCH-MIDAS). This pioneering approach marks a unique innovation. The research employs a diverse array of evaluation methods to validate the model’s robustness and consistently demonstrates its dependable performance. Findings unequivocally reveal the substantial influence of outliers on the model’s predictive capacity, with the AO-GARCH-MIDAS model exhibiting consistent superiority across all evaluation criteria. Additionally, while the GARCH model showcases stronger estimation capabilities compared to the GARCH-MIDAS model, the latter demonstrates heightened predictive prowess. Notably, regarding variable selection, the results underscore the greater predictive informational value inherent in realized volatility over other low-frequency factors.
1 March 2025 in International Journal of ADVANCED AND APPLIED SCIENCES
This study develops an online support system tool based on the Job Demands-Resources (JD-R) model and Engaging Leadership theory to facilitate organizational consensus in evaluating and addressing job characteristics. Using a qualitative research design, including document analysis and the online Delphi method, it constructs a JD-R competency model within the framework of occupational health psychology. The study proposes the classical Delphi method to establish consensus and, in cases of disagreement, employs the Policy Delphi method to explore underlying differences. By integrating these methods into a judgment-based forecasting system, the research provides an effective approach for achieving organizational-level consensus. The proposed system supports Engaging Leadership, mitigates burnout, and enhances employee well-being, offering a practical tool for leadership development and workplace improvement.
5 September 2026 in International Journal of Innovation and Technology Management