Using the two-input Transfer Function (DISO) in Financial Time Series
Main Article Content
Abstract
In most studies of time series analysis, it is noted that most models that are described by one variable may not give accurate future predictions for the series studied, due to lack of attention to their relationship with external indicators of other time series associated with them, as is the case in multiple regression models. Due to the importance of this type of model because it combines the dynamic property of time series with the causal property of multiple regression models, and accordingly, transfer function models were applied with binary inputs and single outputs for monthly observations that include three time series to predict one of the financial indicators (money supply) with two series of inputs (loans). And advances) and (private sector debt). In this research, an alternative method was used to diagnose the ranks of the transfer functions based on the Corner Method. It turned out that the ranks of the transfer functions were Tand T
For the purpose of investigating the accuracy of the model (DISO) matching the data, three other models were applied to the aforementioned data for comparison. Two of them are related to the single-input-single-output (SISO) transfer function models, in addition to the ARIMA model. The application results showed the preference of the (DISO) model because it gave the lowest value for the two criteria (MSE) and (AIC) compared to other models
Article Details
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.