Commodity price prediction using ensembles of neural networks

• Main Author: Hoon, Brian Yong Sheng
• Other Authors: Wang Lipo
• Format: Final Year Project
• Language: English
• Published: 2019
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/77638
• Institution: Nanyang Technological University

The prediction of market prices plays a major role in today’s financial markets. Such prices range from stocks, bonds, estate to commodities such as precious metals. Consequently, forecasting methodologies and techniques have become increasingly vital to the lifeblood of an investor, pushing the need for further research on more effective methods. Hence, silver was chosen as the main subject of this paper due to its volatility. Artificial Neural Network (ANN) and ensembles methods were visited in this paper. Specifically, Backpropagation (BP) and Radial Basis Function (RBF) based models as well as Bootstrap Aggregating (Bagging) and Boosting ensemble methods were evaluated. The main environment utilized for the processing and visualization of data as well as the development and evaluation of ensemble models was MATLAB. It was discovered that among the combinations of neural network and ensemble models, bagging with RBF produces the best prediction results.