Abstract:

The goods expressed as commercial goods are divided into many subgroups, such as industrial metals, precious metals, agricultural products and energy products. The uncertainty that arises with the increase in the price of oil, which is among the high-volume instruments and ranked first in primary energy consumption, causes consumers and producers to change their spending, savings and investment decisions, and potentially to reassign resources. In high-frequency series, the time-changing and accumulating tendency of the gameplay modeling is mostly used by the GARCH (Generalized Otoregressive Conditional Variant) model. However, the Artificial Nervous Networks (ANN), which can also model the curved structure alongside its linear structure, appears to be a good option for the GARCH-type models. In the framework of the study, the structure of the hybrid model (Tip-II) obtained by the inclusion of the multi-variable GARCH type models that are not yet widely used in literature as a system, using the structure of the hybrid model (Tip-II) obtained by the inclusion of the output layer in ANN, the structure of the gameplay in the oil prices from September-1992 and July-2019 periods is studied. Compared to the forecasts obtained by hybrid models, good performance values are achieved by the dynamic conditional correlation model (DCC-MGARCH) and multi-layer detection models (MLP) that belong to the multi-variant GARCH-type model class.

Keywords: