Forecasting Crude Oil Prices: A Comparison between Artificial Neural Networks and Vector Autoregressive Models

Please enable javascript in your browser.

Ramyar, Sepehr | 2015

752 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 47445 (01)
University: Sharif University of Technology
Department: Industrial Engineering
Advisor(s): Kianfar, Farhad
Abstract:
Crude oil is a key element in world economy and the most widely traded form of energy. Therefore, a clear and effective understanding of crude oil price behavior is of great importance for businesses, governments and policy makers. Taking into account the exhaustible nature of crude oil and impact of monetary policy along with other major drivers of crude oil prices, this paper investigates predictability of oil prices using artificial neural networks. A Multilayer Perceptron (MLP) neural network is developed and trained with historical data from 1980 to 2014 and using mean square error (MSE) for testing data, optimal number of hidden layer neurons is determined. Meanwhile, an economic model for crude oil prices is developed and estimated using a vector autoregressive model. Results from the proposed ANN are then compared to those of the Vector Autoregressive model and it is concluded that an MLP neural network can more accurately predict crude oil prices than a VAR model. Ultimately, several scenarios are designed in order to account for plausible conditions of the crude oil market and crude oil price projections are made under each scenario
Keywords:
Forecasting ; Artificial Neural Network ; Vector Autoregressive Model ; Price Forecasting ; Oil Price ; Crude Oil ; Scenario Analysis

No TOC