DSS for Oil Price Prediction Using Machine Learning

Abstract

The oil price affects the economic situation of many countries in the world, therefore, there is always an increased interest. A number of efforts have been made by researchers towards developing efficient methods for forecasting oil prices. In this paper, three types of models for forecasting oil prices were created: Linear Regression, Support Vector Machine (SVM) and Convolutional Neural Network (CNN). The root mean square error and standard error were chosen to estimate the constructed models by quantitative characteristics. For visual analysis the graphs depicting the actual and forecast values were plotted. According to the interpretation of the results to the evaluation criteria of the models, when using the price of Brent oil as input data, the SVM has the best predictive ability. This makes it a good tool for forecasting dynamically changing data of large volumes. Also a model of the decision support system (DSS) architecture, a forecasting subsystem and a forecasting module are designed to show how the results of the study can be used in the work of commodity market traders. Keywords—oil prices, time series, prediction, neural network, support vector machine, machine learning, energy resources, deep learning