Hybridization of Support Vector Regression and Arima Models with Genetic Algorithms for Predicting Crude Oil Price

Abstract

Abstract Crude oil plays a pivotal role in global economics, serving as a crucial raw material for manufacturing and a primary ingredient in transportation gasoline. Accurate forecasting of crude oil prices is essential for various sectors. Conventional statistical and econometric models often struggle with the non-linear and inconsistent nature of crude oil price data, leading to poor prediction performance. In this study, we propose a novel hybrid approach combining Support Vector Regression (SVR), a nonlinear machine learning model, with Autoregressive Integrated Moving Average (ARIMA), a linear econometric model. Genetic Algorithms (GA) was employed to optimize the parameters of both models. Our hybrid models, namely SVRGA_ARIMA and SVRGA_ARIMAGA, outperform individual models such as ARIMA and SVR, as well as the hybrid model SVR_ARIMA, in terms of forecasting accuracy. The proposed hybrid models achieve a significantly lower root mean square error compared to other models. Overall, our findings suggest that the hybrid SVRGA_ARIMA and SVRGA_ARIMAGA models, optimized with GA, offer a reliable framework for forecasting weekly crude oil prices.