Forecasting Daily Crude Oil Prices Using Improved CEEMDAN and Ridge Regression-Based Predictors

Abstract

As one of the leading types of energy, crude oil plays a crucial role in the global economy. Understanding the movement of crude oil prices is very attractive for producers, consumers and even researchers. However, due to its complex features of nonlinearity and nonstationarity, it is a very challenging task to accurately forecasting crude oil prices. Inspired by the well-known framework “decomposition and ensemble” in signal processing and/or time series forecasting, we propose a new approach that integrates the improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN), differential evolution (DE) and several types of ridge regression (RR), namely, ICEEMDAN-DE-RR, for more accurate crude oil price forecasting in this paper. The proposed approach consists of three steps. First, we use the ICEEMDAN to decompose the complex daily crude oil price series into several relatively simple components. Second, ridge regression or kernel ridge regression is employed to forecast each decomposed component. To enhance the accuracy of ridge regression, DE is used to jointly optimize the regularization item, the weights and parameters of each single kernel for each component. Finally, the predicted results of all components are aggregated as the final predicted results. The publicly available West Texas Intermediate (WTI) daily crude oil spot prices are used to validate the performance of the proposed approach. The experimental results indicate that the proposed approach can achieve better performance than some state-of-the-art approaches in terms of several evaluation criteria, demonstrating that the proposed ICEEMDAN-DE-RR is very promising for daily crude oil price forecasting.

Daily Crude Oil Price Forecasting Based on Improved CEEMDAN, SCA, and RVFL: A Case Study in WTI Oil Market

“…The parameters of each method and the parameter ranges of ICEEMDAN and RVFL optimized by SCA in the experiments are listed in Table 1. The values of parameters of EEMD, ICEEMDAN, ARIMA, BPNN, LSSVR and RVFL are from previous literature [18,24].…”

“…To further evaluate the proposed ICEEMDAN-SCA-RVFL model, we compare it with some extant ensemble models using the same framework of "decomposition and ensemble", including ICEEMDAN-DE-RR [24], CEEMD-A&S-SBL [3], and EEMD-APSO-RVM [30]. The experimental results are reported in Table 6, where the best prediction results are shown in bold.…”

“…To cope with that, a "decomposition and ensemble" framework was widely introduced into time series forecasting. The first step of the "decomposition and ensemble" framework is to decompose the complex raw time series into a group of relatively simple components, then a single predictor is introduced to predict each component independently, and finally these predicted values of all components are assembled as the final predicted results [24][25][26][27]. This idea has also been introduced into the field of crude oil price prediction.…”

“…He et al used the heterogeneous features of crude oil price movement to build a novel multivariate EMD-based model in international crude oil markets [29]. Li et al developed a new ensemble model that incorporated ICEEMDAN with ridge regression (RR) for crude oil price prediction [24].…”

Short-Term Load Forecasting with Improved CEEMDAN and GWO-Based Multiple Kernel ELM

“…e main idea of this framework is to decompose the raw energy data series into several simpler components, then handle each component individually, and finally integrate the result from each component as the final forecasting result. is framework is a typical form of the strategy of "divide and conquer" that is widely used in energy price forecasting [22][23][24], wind speed forecasting [25,26], load forecasting [27,28], biosignal processing [29,30], fault diagnosis [31], image processing [32][33][34][35], and so on. ere are many types of decomposition methods that can be applied to decomposing energy time series.…”