Modelling the nexus of macro-economic variables with WTI Crude Oil Price: A Machine Learning Approach

• Published in: 2022 IEEE Region 10 Symposium (TENSYMP) pp. 1 - 6
• Main Authors: Bhagat, Vandana, Sharma, Manjari, Saxena, Anshul
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.07.2022
• Subjects: Analytical models; Consumer Preference Index (CPI); Crude Oil Price; Economic indicators; Feature Engineering; Forecasting; Gross Domestic Product (GDP); Machine learning; Machine learning algorithms; Oils; Predictive Analysis; Reactive power; Time series analysis
• ISSN: 2642-6102
• DOI: 10.1109/TENSYMP54529.2022.9864544

Crude oil price shocks have a significant impact on aggregate macroeconomic indices like GDP, interest rates, investment, inflation, unemployment, and currency rates, according to empirical evidence. Various factors like GDP, CPI, and Gold prices show a considerable impact on the Crude old prices. The correlation analysis between these variables can help the machine learning model to find the highly impacting factor of the target variable. The advanced machine learning algorithms can be used to find the most relevant variable impacting the crude oil price followed by predicting the crude oil price. Time series analysis algorithms can forecast the crude oil prices for the specific period ahead. In the current study it was observed that US dollar and CPI show a high impact on Crude oil prices. The study has implemented six machine learning algorithms out of which the ARIMAX was found to be the most efficient model. VAR and ARIMA models are used to successfully forecast the crude oil prices for the next 5 years. From the current research, a machine learning model has been obtained as an outcome of the study, which will help economists in the future to understand the dynamics of crude oil prices driver and forecast it for the near future.