prediction algorithm

Predicting the price trend of crude oil and its fluctuations has always been one of the challenges facing traders in oil markets. In this regard, in order to make rational predictions about the future, efforts have been made to establish and expand quantitative and qualitative forecasting methods. Therefore, forecasting methods have always been important tools for future researchers. Consequently, this study aims to investigate and evaluate algorithms, especially machine learning algorithms (such as Bayesian networks, artificial neural networks, support vector machines, k-nearest neighbors), for predicting crude oil prices. For this current research, variables such as European Union oil prices and U.S. oil prices from 1990 to 2020 are considered. Python software was used for analysis.The results of this study have shown that, based on the machine learning algorithms used in the first model (Brent crude oil price), algorithms like Linear Regression, Least Angle Regression, Orthogonal Matching Pursuit, and Bayesian Ridge are the best algorithms for predicting the Brent crude oil price. Additionally, the results of machine learning algorithms in the second model (global Brent oil price and West Texas oil price) indicate that Linear Regression, Least Angle Regression, Orthogonal Matching Pursuit, and Bayesian Ridge are the best algorithms for predicting West Texas oil prices.