Exploring machine learning techniques for oil price forecasting: A comparative study of SVM, SMO, and SGD-base models

Abstract

Crude oil prices play a critical role in the global economy, and accurately predicting their future values is of great importance. In this study, we evaluated the performance of three support vector machine (SVM) models - standard SVM, SMO-based SVM, and SGD-based SVM - in predicting crude oil prices based on daily, weekly, and monthly data, aligning with the Industry 4.0 paradigm, to enhance decision-making in the oil and gas sector. The results showed that all three models performed well, with high coefficient of determination R2 values and low MSE values across all versions of the dataset. The SMO and SGD-based models demonstrated particular advantages for larger datasets. Furthermore, SVMs are well-suited for modelling complex financial datasets and can handle non-linear relationships between variables. Our findings suggest that SVMs can provide effective models for predicting crude oil prices, with the SMO-based and SGD-based models showing potential for large datasets.