Ensemble Incremental Random Vector Functional Link Network for Short-term Crude Oil Price Forecasting

Abstract

In this paper, an ensemble incremental learning model composed of Empirical Mode Decomposition (EMD), Random Vector Functional Link network (RVFL) and Incremental RVFL is presented in this work. First of all, EMD is employed to decompose the historical crude oil price time series. Then each sub-signal is modeled by an RVFL model to generate the corresponding forecast IMF value. Finally, the prediction results of all IMFs are combined to formulate an aggregated output for crude oil price. By introducing incremental learning, along with EMD based ensemble methods into RVFL network, the forecasting performance can be significantly improved with respect to both efficiency and accuracy. The crude oil price datasets from West Texas Intermediate (WTI) and Brent oil are used to test the effectiveness of the proposed EMD-Incremental-RVFL method. Simulation results demonstrated attractiveness of the proposed method compared with seven benchmark methods including long short-term memory (LSTM) network, especially based on fast computation speed.