Fuzzy Autoregressive Distributed Lag model-based forecasting

This research aims to be guided decision-makers in future planning by estimating the tendency of data consistently. In this context, it is thought that the integration of the Autoregressive Distributed Lag-ARDL models, gathering the independent factors and their past effects as well as the past trend of the dependent variable, with fuzzy regression methods, would give more realistic results. To prove the correctness of this idea, the Fuzzy-ARDL method has been proposed and tested the superiority of the research on the projection of USAs' annual oil consumption data examined by researchers previously. For this purpose, raw data of crude oil import price, population, gross national domestic production (GDP) per capita, and oil production variables, previously compiled annually, have been considered independent variables. Then the proposed model has been benchmarked with the other promising models from the fuzzy regression literature. As a result, according to various Accuracy Measures values, it has been seen that the proposed model outperforms the other promising models.(c) 2022 Elsevier B.V. All rights reserved.

Automated summary

This research aims to guide decision-makers in future planning by consistently estimating the trend of data. The integration of the Autoregressive Distributed Lag-ARDL models with fuzzy regression methods is believed to provide more realistic results. The proposed Fuzzy-ARDL method outperforms other models in projecting the annual oil consumption data of the USA.