Prediction of Crude Oil Prices using Hybrid Guided Best-So-Far Honey Bees Algorithm-Neural Networks

• Published in: International journal of advanced computer science & applications Vol. 10; no. 5
• Main Authors: Tairan, Nasser, Shah, Habib, Aleryani, Aliya
• Format: Journal Article
• Language: English
• Published: West Yorkshire Science and Information (SAI) Organization Limited 2019
• Subjects: Algorithms; Artificial neural networks; Crude oil; Crude oil prices; Economic development; Heuristic methods; Honey; Neural networks; Predictions; Search algorithms; Swarm intelligence; Topology
• ISSN: 2158-107X; 2156-5570; 2156-5570
• DOI: 10.14569/IJACSA.2019.0100540

The objective of this paper is the use of new hybrid meta-heuristic method called Guided Best-So-Far Honey Bees Inspired Algorithm with Artificial Neural Network (ANN) on the Prediction of Crude Oil Prices of Kingdom of Saudi Arabia (KSA). Very high volatility of crude oil prices is one of the main hurdles for the economic development; therefore, it’s the need of the hour to predict crude oil prices, especially for oil-rich countries such as KSA. Hence, in this paper, we are proposing a hybrid algorithm, named: Guided Best-So-Far Artificial Bee Colony (GBABC) algorithm. The proposed algorithm has been trained and tested with ANN for finding the optimal weight values to increase the exploration and exploitation process with balance quantities to obtain the accurate prediction of crude oil prices. The KSA crude oil prices of the five years 2013 to 2017 have been used to train ANN with different topologies and learning parameters of the proposed method for the prediction of the crude oil prices of the next day. The simulation results have been very promising and encouraging of the proposed algorithm when compared and analyzed with ABC, GABC (Gbest Guided ABC) and Best-So-Far ABC methods for prediction purpose. In most cases, the actual prices and predicted crude oil KSA prices are very close, which were obtained by the proposed GBABC method based on the optimal weight values of ANN and minimum prediction error.