Evolutionary Neural Network model for West Texas Intermediate crude oil price prediction Haruna Chiroma a , Sameem Abdulkareem a , Tutut Herawan b,⇑ a Department of Artificial Intelligence, Faculty of Computer Science and Information Technology, University of Malaya, 50603 Pantai Valley, Kuala Lumpur, Malaysia b Department of Information System, Faculty of Computer Science and Information Technology, University of Malaya, 50603 Pantai Valley, Kuala Lumpur, Malaysia highlights  We propose approach for the prediction of the WTI crude oil price.  The values predicted by the proposed method and actual once are statistically equal.  The proposed method indicated performance improvement over existing results. article info Article history: Received 27 August 2014 Received in revised form 24 November 2014 Accepted 21 December 2014 Available online 17 January 2015 Keywords: Genetic Algorithm Neural Network West Texas Intermediate crude oil price Backpropagation algorithms abstract This paper proposes an alternative approach based on a genetic algorithm and neural network (GA–NN) for the prediction of the West Texas Intermediate (WTI) crude oil price. Comparative simulation results suggested that the proposed GA–NN approach is better than the baseline algorithms in terms of predic- tion accuracy and computational efficiency. Mann–Whitney test results indicated that the WTI crude oil price predicted by the proposed GA–NN and the observed price are statistically equal. Further compari- son of the proposed GA–NN with previous studies indicated performance improvement over existing results. The proposed model can be useful in the formulation of policies related to international crude oil price estimations, development plans and industrial production. Ó 2014 Elsevier Ltd. All rights reserved. 1. Introduction The price of crude oil was $11 per barrel for 25 years, but from February 1999 to September 2000, the price raised to a peak of almost $35 per barrel. In November 2001, the price of crude oil fell significantly, thereby slowing world economic activities [1]. In July 2008, the price of West Texas Intermediate (WTI) crude oil climbed a peak value of more than US $147 per barrel, reaching a record high. The global financial crises that destabilised the world econ- omy triggered the WTI crude oil price to crash to approximately US $30 per barrel in the early first quarter of 2009. In 2011, the price of oil regained value again and shot up to US $100 per barrel. The price of international crude oil is known to exhibit a complex behaviour, and its price dynamic is affected by many factors [2]. Therefore, the WTI oil market has attracted attention from researchers in recent times. The crude oil price has been part of the decision-making process for development and production in industries, as well as government short- and long-term planning, export policy and national reserves. Therefore, its accurate predic- tion has become a critical issue to both governments and industries for accurate decision making [3]. The prediction of crude oil price is an active area of research in the literature, in an effort to develop a reliable system that can predict its behaviour. Hence, it is impor- tant to provide decision makers with predictions of future occur- rences of its patterns so that they can be used for national and international development plans and reduce the hardship typically imposed by the hike of the crude oil price. There are several studies in the literature for the prediction of the crude oil price. For example, Reza and Ahmadi [4] used Genetic Algorithm (GA) for selecting Neural Network (NN) hidden layer neurons, activation function, and the number of layers including connections. Then, Levenberg – Marquardt backpropagation (LMBP) algorithm was used to train the network and optimised the NN weights. He et al. [5] adopted feed-forward NN (FFNN) due to its computational efficiency over other NN architecture such as recurrent NN. The FFNN is trained with LMBP to build ensemble models to enhance the forecast accuracy of the crude oil price. Jammazi and Aloui [6] built a hybrid artificial intelligence model http://dx.doi.org/10.1016/j.apenergy.2014.12.045 0306-2619/Ó 2014 Elsevier Ltd. All rights reserved. ⇑ Corresponding author. Tel.: +60 142723760; fax: +60 379579249. E-mail addresses: hchiroma@acm.org (H. Chiroma), sameem@um.edu.my (S. Abdulkareem), tutut@um.edu.my (T. Herawan). Applied Energy 142 (2015) 266–273 Contents lists available at ScienceDirect Applied Energy journal homepage: www.elsevier.com/locate/apenergy