Multivariate EMD-Based Modeling and Forecasting of Crude Oil Price

• Main Authors: Kaijian He, Rui Zha, Jun Wu, Kin Keung Lai
• Format: Article
• Language: English
• Published: MDPI AG 2016-04-01
• Series: Sustainability
• Subjects: empirical mode decomposition (EMD); multivariate EMD analysis; crude oil price forecasting; time delay embedding; multiscale analysis; ARMA model
• Online Access: http://www.mdpi.com/2071-1050/8/4/387

Recent empirical studies reveal evidence of the co-existence of heterogeneous data characteristics distinguishable by time scale in the movement crude oil prices. In this paper we propose a new multivariate Empirical Mode Decomposition (EMD)-based model to take advantage of these heterogeneous characteristics of the price movement and model them in the crude oil markets. Empirical studies in benchmark crude oil markets confirm that more diverse heterogeneous data characteristics can be revealed and modeled in the projected time delayed domain. The proposed model demonstrates the superior performance compared to the benchmark models.