A Comparative Study on Novel Hybrid Approaches Based on CEEMDAN, Random Forest, Deep Learning Methods for Predicting Daily Wind Speed

In the realm of power systems, short-term electric load forecasting is pivotal for ensuring supply–demand balance, optimizing generation planning, reducing operational costs, and maintaining grid stability. Short-term load curves are characteristically coarse, revealing high-frequency data upon decomposition that exhibit pronounced non-linearity and significant noise, complicating efforts to enhance forecasting precision. To address these challenges, this study introduces an innovative model. This model employs complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) to bifurcate the original load data into low- and high-frequency components. For the smoother low-frequency data, a temporal convolutional network (TCN) is utilized, whereas the high-frequency components, which encapsulate detailed load history information yet suffer from a lower fitting accuracy, are processed using an enhanced soft thresholding TCN (SF-TCN) optimized with the slime mould algorithm (SMA). Experimental tests of this methodology on load forecasts for the forthcoming 24 h across all seasons have demonstrated its superior forecasting accuracy compared to that of non-decomposed models, such as support vector regression (SVR), recurrent neural network (RNN), gated recurrent unit (GRU), long short-term memory (LSTM), convolutional neural network-LSTM (CNN-LSTM), TCN, Informer, and decomposed models, including CEEMDAN-TCN and CEEMDAN-TCN-SMA.

Accurate prediction of Global Navigation Satellite System-derived precipitable water vapor (GNSS-PWV), which is a crucial indicator for climate change monitoring, holds significant scientific value for climate disaster prevention and mitigation. In the study of GNSS-PWV prediction, the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) algorithm within a decomposition–integration framework effectively addresses the non-stationarity and complexity of PWV sequences, enhancing prediction accuracy. However, residual noise and pseudo-modes from decomposition can distort signals, reducing the predictor system’s reliability. Additionally, independent modeling of all decomposed components decreases computational efficiency. To address these challenges, this paper proposes a hybrid model combining the improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN), autoregressive integrated moving average (ARIMA), and long short-term memory (LSTM) networks. Enhanced by local mean optimization and adaptive noise regulation, the ICEEMDAN algorithm effectively suppresses pseudo-modes and minimizes residual noise, enabling its decomposed intrinsic mode functions (IMFs) to more accurately capture the multi-scale features of GNSS-PWV. Sample entropy (SE) is used to quantify the complexity of IMFs, and components with similar entropy values are reconstructed into the following three sub-sequences: high-frequency, low-frequency, and trend. This process significantly reduces modeling complexity and improves computational efficiency. We propose different modeling strategies tailored to the dynamics of various subsequences. For the nonlinear and non-stationary high-frequency components, the LSTM network is used to effectively capture their complex patterns. The LSTM’s gating mechanism and memory cell design proficiently address the long-term dependency issue. For the stationary and weakly nonlinear low-frequency and trend components, linear patterns are extracted using ARIMA. Differencing eliminates trends and moving average operations capture random fluctuations, effectively addressing periodicity and trends in the time series. Finally, the prediction results of the three components are linearly combined to obtain the final prediction value. To validate the model performance, experiments were conducted using measured GNSS-PWV data from several stations in Hong Kong. The results demonstrate that the proposed model reduces the root mean square error by 56.81%, 37.91%, and 13.58% at the 1 h scale compared to the LSTM, EMD-LSTM, and ICEEMDAN-SE-LSTM benchmark models, respectively. Furthermore, it exhibits strong robustness in cross-month forecasts (accounting for seasonal influences) and multi-step predictions over the 1–6 h period. By improving the accuracy and efficiency of PWV predictions, this model provides reliable technical support for the real-time monitoring and early warning of extreme weather events in Hong Kong while offering a universal methodological reference for multi-scale modeling of geophysical parameters.

Financial time series forecasting model based on CEEMDAN and LSTM

Accurate forecasting of hourly water demand is essential for effective and sustainable operation, and the cost-effective management of water distribution networks. Unlike monthly or yearly water demand, hourly water demand has more fluctuations and is easily affected by short-term abnormal events. An effective preprocessing method is needed to capture the hourly water demand patterns and eliminate the interference of abnormal data. In this study, an innovative preprocessing framework, including a novel local outlier detection and correction method Isolation Forest (IF), an adaptive signal decomposition technique Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN), and basic forecasting models have been developed. In order to compare a promising deep learning method Gated Recurrent Unit (GRU) as a basic forecasting model with the conventional forecasting models, Support Vector Regression (SVR) and Artificial Neural Network (ANN) have been used. The results show that the proposed hybrid method can utilize the complementary advantages of the preprocessing methods to improve the accuracy of the forecasting models. The root-mean-square error of the SVR, ANN, and GRU models has been reduced by 57.5%, 27.8%, and 30.0%, respectively. Further, the GRU-based models developed in this study are superior to the other models, and the IF-CEEMDAN-GRU model has the highest accuracy. Hence, it is promising that this preprocessing framework can improve the performance of the water demand forecasting models.

The performance of tissue imaging is improved by tissue harmonic imaging (THI) to expand and enhance the range and level of clinical diagnosis for many diseases. The separation based on High-pass filtering (S_HPF) is a commonly used method for extracting harmonic components from ultrasonic echo signals. However, the cut-off frequency, order and algorithm to realize a high-pass filter have a great influence on the separation accuracy of harmonic signals. In present study, a novel separation method (S_CEEMDAN) based on the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) algorithm is proposed for adaptively separating harmonic components from ultrasonic echo signals. First, we calculate the ensemble size of CEEMDAN adaptively based on the added noise level, and draw a sum of intrinsic mode functions (IMFs) from the ultrasonic echo signals by CEEMDAN. Then, the spectrum of each IMF is computed and evaluated, and the IMFs containing both fundamental and harmonic components are further decomposed by using the CEEMDAN algorithm. This separation process is end until all of IMFs have been divided into either fundamental or harmonic categories. Finally, the corresponding fundamental and harmonic echo signals are yielded by accumulating separately these two categories. In experiments, simulated ultrasonic echo signals with a center frequency of 3.5MHz are separated by the proposed S_CEEMDAN method, and the results are compared with those processed by S_HPF. The edge of the harmonic image by S_CEEMDAN marginally better defined than S_HPF. The indices for the harmonic signals separated by S_CEEMDAN and S_HPF, respectively, are as follows: the center frequencies of 6.66 MHz and 6.65 MHz, 3 dB bandwidths of 1.04 MHz and 1.05 MHz, 6 dB bandwidths of 1.59 MHz and 1.55 MHz, signal-to-noise ratios (SNRs) of 14.00 and 13.63, and image contrasts of 7.01 and 7.06. In conclusion, due to good adaptive characteristics, and lower reconstruction errors, the proposed S_CEEMD method is superior to S_HPF in the performance of spectral accuracy and harmonic imaging. This method could be potentially alternative to the current method for the ultrasonic harmonic separation.

Long-short term memory (LSTM) is a state-of-art and widely used model to forecast financial time series. However, primitive LSTM networks do not perform well due to over-fitting problems of the deep learning model and nonlinear and non-stationary characteristics of financial time series data. In addition, complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) is an outstanding data frequency decomposition technique that can decompose original time series into several intrinsic mode functions and a residue. Thus, this paper proposed a novel hybrid network CEEMDAN-LSTM-BN based on LSTM. Specifically, to avoid over-fitting, the modified LSTM-BN network consists of two LSTM layers, two Batch Normalization (BN) layers following each LSTM layer, and a dropout layer. Each of the intrinsic mode functions and the residue would be processed by CEEMDAN-LSTM-BN and the final prediction results are obtained by reconstructing each predictive series. The advantages of the proposed CEEMDANLSTM-BN networks are verified by comparing them to primitive LSTM, other hybrid models, and some famous machine learning models. Moreover, the robustness of the networks is assessed by numerical experiments on different stock indices datasets.

Variation of the Kuroshio path south of Japan has an important impact on weather, climate, and ecosystems due to its distinct features. Motivated by the ever-popular deep learning methods using neural network architectures in areas where more accurate reference data for oceanographic observations and reanalysis are available, we build four deep learning models based on the long short-term memory (LSTM) neural network, combined with the empirical orthogonal function (EOF) and complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN), namely, the LSTM, EOF–LSTM, CEEMDAN–LSTM, and EOF–CEEMDAN–LSTM. Using these models, we conduct long-range predictions (120 days) of the Kuroshio path south of Japan based on 50-yr ocean reanalysis and nearly 15 years of satellite altimeter data. We show that the EOF–CEEMDAN–LSTM performs the best among the four models, by attaining approximately 0.739 anomaly correlation coefficient and 0.399° root-mean-square error for the 120-day prediction of the Kuroshio path south of Japan. The hindcasts of the EOF–CEEMDAN–LSTM are successful in reproducing the observed formation and decay of the Kuroshio large meander during 2004/05, and the formation of the latest large meander in 2017. Finally, we present predictions of the Kuroshio path south of Japan at 120-day lead time, which suggest that the Kuroshio will remain in the state of the large meander until November 2022.

ABSTRACTThis study aims to forecast the S&P 500 ESG index using the mixture model of complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and long short‐term memory (LSTM) prediction models. CEEMDAN enables decomposing the index's original return series into different intrinsic mode functions (IMFs) and a residual series. The decomposed IMFs are then regrouped into aggregate series depicting high frequency and medium frequency, while the residual series represent the trend component. LSTM algorithm is used on the aggregated series to obtain predicted values of the same. The study compares different prediction algorithms to identify their performance and explore the predictive power of the hybrid models.

Prediction of dissolved oxygen in a fishery pond based on gated recurrent unit (GRU)

Numerous studies show that it is reasonable and effective to apply decomposition technology to deal with the complex carbon price series. However, the existing research ignores the residual term containing complex information after applying single decomposition technique. Considering the demand for higher accuracy of the carbon price series prediction and following the existing research path, this paper proposes a new hybrid prediction model VMD-CEEMDAN-LSSVM-LSTM, which combines a new quadratic decomposition technique with the optimized long short term memory (LSTM). In the decomposition part of the hybrid model, the original carbon price series is processed by variational mode decomposition (VMD), and then the residual term obtained by decomposition is further decomposed by complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN). In the prediction part of the hybrid model, least squares support vector machine (LSSVM) is introduced, and LSSVM-LSTM model is constructed to predict the components obtained by decomposition. The empirical research of this paper selects two different case data from the European Union emissions trading system (EU ETS) as samples. Taking the results of Case Ⅰ in the 1-step ahead forecasting scenario as an example, the prediction evaluation indexes eMAPE, eRMSE and R2 of the VMD-CEEMDAN-LSSVM-LSTM hybrid model constructed in this paper are 0.3087, 0.0921 and 0.9987 respectively, which are significantly better than other benchmark models. The empirical results confirm the superiority and robustness of the hybrid model proposed in this paper for carbon price forecasting.

Accurate power load forecasting plays an important role in smart grid analysis. To improve the accuracy of forecasting through the three-level “decomposition–optimization–prediction” innovation, this study proposes a prediction model that integrates complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN), the improved sparrow search algorithm (ISSA), a convolutional neural network (CNN), and bidirectional long short-term memory (BiLSTM). A series of simpler intrinsic mode functions (IMFs) with different frequency characteristics can be decomposed by CEEMDAN from data, then each IMF is reconstructed based on calculating the sample entropy of each IMF. The ISSA introduces three significant enhancements over the standard sparrow search algorithm (SSA), including that the initial distribution of the population is determined by the optimal point set, the position of the discoverer is updated by the golden sine strategy, and the random walk of the population is enhanced by the Lévy flight strategy. By the optimization of the ISSA to the parameters of the CNN-BiLSTM model, integrating the prediction results of the reconstructed IMFs in the sub-models can obtain the final prediction result of the data. Through the performance indexes of the designed prediction model, the application case results show that the proposed combined prediction model has a smaller prediction error and higher prediction accuracy than the eight comparison models.

Accurate photovoltaic (PV) power prediction technology plays a crucial role in effectively addressing the challenges posed by the integration of large-scale PV systems into the grid. In this article, we propose a novel PV power combination prediction model based on complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) in conjunction with a hybrid neural network. To mitigate the influence of strong fluctuations in PV power on prediction outcomes, we employ CEEMDAN to decompose the PV data into several subsequences. Subsequently, sample entropy (SE) is used to quantify the complexity of each subsequence. Subsequences with similar SE values are then restructured to reduce computational load. Moreover, to overcome the limitations of a single neural network in capturing historical data features of PV power, a hybrid sequential convolutional neural network-gate recurrent unit (CNN-GRU) neural network is proposed. The effectiveness of our proposed model is validated through case studies involving PV power stations in two regions. To provide a comprehensive assessment, we conduct comparative validation by building and evaluating alternative models, including long-short term memory (LSTM), GRU, CEEMDAN-LSTM, CEEMDAN-GRU, and CNN-GRU. The results unequivocally demonstrate that the model presented in this article exhibits exceptional prediction performance, characterized by high accuracy and robust generalization.

Warships play an important role in the modern sea battlefield. Research on the line spectrum features of warship radio noise signals is helpful to realize the classification and recognition of different types of warships, and provides critical information for sea battlefield. In this paper, we proposed a novel linear spectrum frequency feature extraction technique for warship radio noise based on complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN), duffing chaotic oscillator (DCO), and weighted-permutation entropy (W-PE). The proposed linear spectrum frequency feature extraction technique, named CEEMDAN-DCO-W-PE has the following advantages in comparison with other linear spectrum frequency feature extraction techniques; (i) as an adaptive data-driven algorithm, CEEMDAN has more accurate and more reliable decomposition performance than empirical mode decomposition (EMD) and ensemble EMD (EEMD), and there is no need for presetting parameters, such as decomposition level and basis function; (ii) DCO can detect the linear spectrum of narrow band periodical warship signals by way of utilizing its properties of sensitivity for weak periodical signals and the immunity for noise; and (iii) W-PE is used in underwater acoustic signal feature extraction for the first time, and compared with traditional permutation entropy (PE), W-PE increases amplitude information to some extent. Firstly, warship radio noise signals are decomposed into some intrinsic mode functions (IMFs) from high frequency to low frequency by CEEMDAN. Then, DCO is used to detect linear spectrum of low-frequency IMFs. Finally, we can determine the linear spectrum frequency of low-frequency IMFs using W-PE. The experimental results show that the proposed technique can accurately extract the line spectrum frequency of the simulation signals, and has a higher classification and recognition rate than the traditional techniques for real warship radio noise signals.

To address the challenge of reduced prediction accuracy caused by capacity regeneration during the use of lithium-ion batteries, this study proposes an RUL (remaining useful life) prediction method based on mode decomposition and an enhanced Informer-LSTM hybrid model. The capacity is selected as the health indicator, and the CEEMDAN (complete ensemble empirical mode decomposition with adaptive noise) algorithm is employed to decompose the capacity sequence into high-frequency and low-frequency components. The high-frequency components are further decomposed and predicted using the Informer model, while the low-frequency components are predicted with an LSTM (long short-term memory) network. Pearson correlation coefficients between each component and the original sequence are calculated to determine fusion weights. The final RUL prediction is obtained through weighted integration of the individual predictions. Experimental validation on publicly available NASA and CALCE (Center for Advanced Life Cycle Engineering) battery datasets demonstrates that the proposed method achieves an average fitting accuracy of approximately 99%, with MAE (mean absolute error) below 0.02. Additionally, both MAPE (mean absolute percentage error) and RMSE (root-mean-square error) remain at low levels, indicating improvements in prediction precision.

The stock index is an important indicator to measure stock market fluctuation, with a guiding role for investors’ decision-making, thus being the object of much research. However, the stock market is affected by uncertainty and volatility, making accurate prediction a challenging task. We propose a new stock index forecasting model based on time series decomposition and a hybrid model. Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) decomposes the stock index into a series of Intrinsic Mode Functions (IMFs) with different feature scales and trend term. The Augmented Dickey Fuller (ADF) method judges the stability of each IMFs and trend term. The Autoregressive Moving Average (ARMA) model is used on stationary time series, and a Long Short-Term Memory (LSTM) model extracts abstract features of unstable time series. The predicted results of each time sequence are reconstructed to obtain the final predicted value. Experiments are conducted on four stock index time series, and the results show that the prediction of the proposed model is closer to the real value than that of seven reference models, and has a good quantitative investment reference value.