Enhancing Forecasting Accuracy of Financial Time Series by Hybrid VAR and Transfer Function Models

Our study suggests an approach that integrates vector autoregressive and transfer function models to enhance the modeling and forecasting of financial time series. All variables are endogenous, which may limit their ability to capture the influence of external factors in Vector Autoregressive models. In contrast, Transfer function models are particularly useful for incorporating exogenous inputs and exploring the direct effects of these variables on the system. This study suggests a hybrid modeling framework that enables a combined assessment of internal dynamics and external influences on macroeconomic variables by integrating the strengths of both approaches. The study concluded that the Transfer Function Models can improve forecasting accuracy in macroeconomic time series when combined with the Vector Autoregressive model. The Transfer Function Model incorporates the effects of exogenous variables into the hybrid approach, while the Vector Autoregressive model captures the relationships among endogenous variables. The results indicate that the proposed hybrid approach is highly effective and has the advantages of both Transfer Function and Vector Autoregressive Models. These results can be applied to enhance the accuracy of time series, especially financial ones.