Ensemble Empirical Mode Decomposition Algorithm Research Articles

Network traffic prediction model based on ensemble empirical mode decomposition and multiple models

SummaryIn order to improve the prediction accuracy of network traffic, a novel prediction model based on ensemble empirical mode decomposition and multiple models is proposed. In this study, ensemble empirical mode decomposition algorithm is introduced to decompose the network traffic and get several components. Approximate entropy is introduced to judge the complexity of each component. According to the results of approximate entropy, echo state network is selected to predict high complexity components, support vector machine is used to predict medium complexity components, and autoregressive integrated moving average model is introduced to predict low complexity components. The advantages of each prediction model are used to predict the appropriate component. The final prediction results are obtained by adding the predicted values of each component. In order to solve the problem that the prediction performance of support vector machine and echo state network is affected by their parameters, an improved whale optimization algorithm is proposed to optimize the parameters of the model. Meanwhile, the calculation results of approximate entropy show that compared with the original network traffic, the complexity of each component obtained by ensemble empirical mode decomposition is reduced, which reduces the complexity of modeling. Three network traffic datasets with sampling periods of 10 ms, 1 s, and 10 min are collected. Compared with the other five state‐of‐the‐art prediction models, the case study results show that the proposed prediction model has better prediction accuracy and excellent statistical performance indicators.

A Short-Term Wind Powerprediction Model Based on CEEMD and WOA-KELM

Effective short-term wind power prediction is crucial to the optimal dispatching, stability, and operation cost control of a power system. In order to deal with the intermittent and fluctuating characteristics of wind power timing series signals, a hybrid forecasting model is proposed, based on Complementary Ensemble Empirical Mode Decomposition (CEEMD) and Whale Optimization Algorithm (WOA)- Kernel Extreme Learning Machine (KELM), to predict short-term wind power. Firstly, the non-stationary wind power time series is decomposed into a series of relatively stationary components by CEEMD. Then, the components are used as the training set for the KELM prediction model, in which the initial values and thresholds are optimized by WOA. Finally, the predicted output values of each component are superimposed, to obtain the final prediction of the wind power values. The experimental results show that the proposed prediction method can reduce the complexity of the prediction with a small reconstruction error. Furthermore, performance is greater, in terms of prediction accuracy and stability, with lower computational cost than other benchmark models.

A Fusion Method Based on EEMD, Pearson Correlation Analysis, Improved LSTM, and Gaussian Function-Trust Region Algorithm for State of Health Prediction of Lithium-Ion Batteries

Abstract Aiming at the phenomenon of battery capacity regeneration, which leads to inaccurate prediction of lithium-ion battery state of health (SOH), a new fusion method based on ensemble empirical mode decomposition (EEMD), Pearson correlation analysis (PCA), and improved long short-term memory (LSTM) network and Gaussian function-trust region (GS-TR) algorithm is introduced to predict battery SOH. First, the EEMD method is adopted to process the battery SOH data to eliminate the impact of capacity recovery. Second, the decomposed data signals are classified by the PCA method, and the signals classified as high frequency and low frequency are respectively predicted by the improved LSTM algorithm and the GS-TR algorithm. Finally, the prediction results of the improved LSTM and GS-TR algorithms are integrated. The proposed fusion method avoids the complexity of the hybrid neural network model and improves the prediction efficiency. Based on the average results of the three data sets from NASA, the RMSE result of the proposed algorithm is reduced by 9.56% compared with the improved LSTM with the EEMD algorithm and 37.57% compared with the improved LSTM without the EEMD algorithm. The results show that the proposed method has higher adaptability and prediction accuracy.

A hybrid deep learning architecture for wind power prediction based on bi-attention mechanism and crisscross optimization

Accurate wind power forecasting is of great significance for power system operation. In this study, a triple-stage multi-step wind power forecasting approach is proposed by applying attention-based deep residual gated recurrent unit (GRU) network combined with ensemble empirical mode decomposition (EEMD) and crisscross optimization algorithm (CSO). In the data processing stage, the EEMD is used to decompose the wind power/speed time series and a bi-attention mechanism (BA) is applied to enhance the sensitivity of model to the important information from both time and feature dimension. In the prediction stage, a series-connected deep learning model called RGRU consisting of the residual network and GRU is proposed as the forecasting model, aiming to make full use of extracting the static and dynamic coupling relationship among the input features. In the retraining-stage, a high-performance CSO algorithm is adopted to further optimize the fully-connected layer of RGRU model so as to improve the generalization ability of the model. The proposed method is validated on a wind farm located in Spain and the experimental results demonstrate that the proposed hybrid model has significant advantage over other state-of-the-art models involved in this study in terms of prediction accuracy and stability.

A Short-Term Prediction Model of PM2.5 Concentration Based on Deep Learning and Mode Decomposition Methods

Based on a set of deep learning and mode decomposition methods, a short-term prediction model for PM2.5 concentration for Beijing city is established in this paper. An ensemble empirical mode decomposition (EEMD) algorithm is first used to decompose the original PM2.5 timeseries to several high- to low-frequency intrinsic mode functions (IMFs). Each IMF component is then trained and predicted by a combination of three neural networks: back propagation network (BP), long short-term memory network (LSTM), and a hybrid network of a convolutional neural network (CNN) + LSTM. The results showed that both BP and LSTM are able to fit the low-frequency IMFs very well, and the total prediction errors of the summation of all IMFs are remarkably reduced from 21 g/m3 in the single BP model to 4.8 g/m3 in the EEMD + BP model. Spatial information from 143 stations surrounding Beijing city is extracted by CNN, which is then used to train the CNN+LSTM. It is found that, under extreme weather conditions of PM2.5 < 35 g/m3 and PM2.5 > 150 g/m3, the prediction errors of the CNN + LSTM model are improved by ~30% compared to the single LSTM model. However, the prediction of the very high-frequency IMF mode (IMF-1) remains a challenge for all neural networks, which might be due to microphysical turbulences and chaotic processes that cannot be resolved by the above-mentioned neural networks based on variable–variable relationship.

Developing a Novel Water Quality Prediction Model for a South African Aquaculture Farm

Providing an accurate prediction of water quality parameters for improved water quality management is a topical issue in the aquaculture industry. Conventional prediction methods have shown different challenges like a poor generalization, poor prediction accuracy, and high time complexity. Aiming at these challenges, a novel hybrid prediction model with ensemble empirical mode decomposition (EEMD) and deep learning (DL) long-short term memory (LSTM) neural network is proposed in this paper. In this innovative hybrid EEMD-DL-LSTM model, firstly, the integrity of the datasets is enhanced by applying moving average filtering and linear interpolation techniques of water quality parameter datasets pre-treatment. Secondly, the measured real sensor water quality parameters dataset is decomposed with the aid of the EEMD algorithm into disparate IMFs and a corresponding residual item. Thirdly, a multi-feature selection process is applied to make a careful selection of a strongly correlated group of IMFs with the measured real water quality parameter datasets and integrate them as inputs to the DL-LSTM neural network. The presented model is built on water quality sensor data collected from an Abalone farm in South Africa. The performance of the novel hybrid prediction model is validated by comparing the results against the real datasets. To measure the overall accuracy of the novel hybrid prediction model, different statistical indices, namely the Mean Absolute Error (MAE), Mean Square Error (MSE), Root Mean Square Error (RMSE), and Mean Absolute Percentage Error (MAPE), are used.

A framework combining acoustic features extraction method and random forest algorithm for gas pipeline leak detection and classification

Monitoring the operation status of the gas pipeline is of great significance to ensure the safe and stable operation of the pipeline. A new framework combining the acoustic features extraction method and Random Forest (RF) algorithm is proposed for gas pipeline leak detection and classification under the strong background noise. Acoustic signal has the advantages of non-contact measurement and insensitive to structural resonances, but it is easy to collect signals with low Signal-to-Noise Ratio (SNR) under strong background noise. To improve SNR and extract useful acoustic features, the acoustic features extraction method based on Ensemble Empirical Mode Decomposition (EEMD) and Correlation Coefficient (CC) analysis is proposed. CC analysis is used to select appropriate Intrinsic Mode Functions (IMFs), which are decomposed by EEMD algorithm. For each selected IMF, Mel-Frequency Cepstral Coefficients (MFCC), time-domain features, waveform features are extracted to form a feature matrix. Because RF has the characteristic of a small training set requirement, which provides a solution to the difficulty of obtaining enough leak acoustic signals in the industrial environment, RF algorithm is employed for leak detection and classification. Experiments were conducted to verify the effectiveness of the proposed framework. When the sampling distance is set to 1 m and 6 m, the average accuracies of RF with extracted acoustic features are higher than the results with all features. The proposed framework is general and can be applied to various acoustic-based industrial equipment condition monitoring challenges.

Small Floating Target Detection Method Based on Chaotic Long Short-Term Memory Network

In order for the detection ability of floating small targets in sea clutter to be improved, on the basis of the complete ensemble empirical mode decomposition (CEEMD) algorithm, the high-frequency parts and low-frequency parts are determined by the energy proportion of the intrinsic mode function (IMF); the high-frequency part is denoised by wavelet packet transform (WPT), whereas the denoised high-frequency IMFs and low-frequency IMFs reconstruct the pure sea clutter signal together. According to the chaotic characteristics of sea clutter, we proposed an adaptive training timesteps strategy. The training timesteps of network were determined by the width of embedded window, and the chaotic long short-term memory network detection was designed. The sea clutter signals after denoising were predicted by chaotic long short-term memory (LSTM) network, and small target signals were detected from the prediction errors. The experimental results showed that the CEEMD-WPT algorithm was consistent with the target distribution characteristics of sea clutter, and the denoising performance was improved by 33.6% on average. The proposed chaotic long- and short-term memory network, which determines the training step length according to the width of embedded window, is a new detection method that can accurately detect small targets submerged in the background of sea clutter.

Improved short-term prediction of significant wave height by decomposing deterministic and stochastic components

Significant wave height prediction for the following hours is a necessity for the planning and operation of wave energy devices. For a site-specific and short-term prediction, classical numerical wave forecasting methods may not be justified as exhaustive climatological data and huge computational power are needed. In this paper, a combination of a decomposition approach and long short-term memory network was presented to forecast the significant wave heights. An improved version of complete ensemble empirical mode decomposition algorithm and recurrence quantification analysis were applied to separate the original time series into deterministic and stochastic components. Each decomposed series was forecasted by the long short-term memory network and the final predicted significant wave heights were obtained by integrating the deterministic and stochastic predictions. Wave data measured at three buoy stations along the eastern coast of the United States were utilized to verify the hybrid model. The performance of the proposed method in three different wave height ranges was evaluated. The results suggested that the hybrid model outperformed the stand-alone long short-term memory network adjusted on the unseparated signal; in particular, for longer lead times and larger wave heights.

Radar Life Signal Processing Algorithm Based on Improved EEMD

Non-contact measurement of human vital signs is of great significance to the construction of smart homes and hospitals. The vital signal extracted from the FMCW radar contains a lot of noise. In order to reduce the noise impact and improve the life signal detection results accuracy, Ensemble Empirical Mode Decomposition (EEMD) algorithm is studied, and a new method based on the combination of EEMD and Translation Invariant Wavelet Transform (TIWT) is proposed. This method firstly decomposes the original signal using EEMD to obtain several Intrinsic Mode Function (IMF) components, and then changes the arrangement order of IMF by Translation Invariant Wavelet Transform to change the position of the singularity in the whole signal to reduce or eliminate the oscillations. Finally, signal screening is completed and life signals are reconstructed. Experiments show that, the proposed method has better detection accuracy in detecting vital signs collected by radar.

Comparison of Forecasting Models for Real-Time Monitoring of Water Quality Parameters Based on Hybrid Deep Learning Neural Networks

Accurate real-time water quality prediction is of great significance for local environmental managers to deal with upcoming events and emergencies to develop best management practices. In this study, the performances in real-time water quality forecasting based on different deep learning (DL) models with different input data pre-processing methods were compared. There were three popular DL models concerned, including the convolutional neural network (CNN), long short-term memory neural network (LSTM), and hybrid CNN–LSTM. Two types of input data were applied, including the original one-dimensional time series and the two-dimensional grey image based on the complete ensemble empirical mode decomposition algorithm with adaptive noise (CEEMDAN) decomposition. Each type of input data was used in each DL model to forecast the real-time monitoring water quality parameters of dissolved oxygen (DO) and total nitrogen (TN). The results showed that (1) the performances of CNN–LSTM were superior to the standalone model CNN and LSTM; (2) the models used CEEMDAN-based input data performed much better than the models used the original input data, while the improvements for non-periodic parameter TN were much greater than that for periodic parameter DO; and (3) the model accuracies gradually decreased with the increase of prediction steps, while the original input data decayed faster than the CEEMDAN-based input data and the non-periodic parameter TN decayed faster than the periodic parameter DO. Overall, the input data preprocessed by the CEEMDAN method could effectively improve the forecasting performances of deep learning models, and this improvement was especially significant for non-periodic parameters of TN.

Reliability of the prediction model for landslide displacement with step-like behavior

Based on the machine learning algorithms, prediction models for landslide displacement with step-like behavior in the reservoir area were established for landslides prevention and reduction; these models could predict a given test set very well. However, due to the length and the sequence of the training set in prediction models, the predictive ability of these prediction models could not be evaluated accurately if just validated with a given test set. To solve the problem, a hybrid reliability model was proposed. Complimentary ensemble empirical mode decomposition (CEEMD) algorithm was used to decompose the accumulated displacement into the trend displacement and the periodic displacement firstly. The Gauss function was used to predict the trend displacement, and the random forest (RF) and back propagation neural network (BPNN) algorithms were employed to predict the periodic displacement. Furthermore, a novel performance function for the reliability analysis of the displacement prediction model was derived to address failure probabilities in different cases. The Baijiabao landslide in the Three Gorges Reservoir Area was taken as an example for reliability analysis of the prediction model for landslide displacement with step-like behavior, and the predictive ability of the CEEMD-RF model and the CEEMD-BPNN model were compared. The results indicated that the CEEMD-RF model and CEEMD-BPNN model both can accurately predict the accumulated displacement of the given test set; the predictive value obtained with the CEEMD-RF model or the CEEMD-BPNN model showed uncertainties of the prediction model for landslide displacement and the predictive ability of the CEEMD-RF model was more reliable than the CEEMD-BPNN model under different cases. The failure probability proposed in the paper could evaluate the predictive ability of the model more accurately and comprehensively compared with the existing assessment indices.

Short-term photovoltaic power prediction on modal reconstruction: A novel hybrid model approach

The contribution is to propose a novel hybrid model based on modal reconstruction to predict short-term photovoltaic (PV) power. PV power generation large-scale grid connection causes the impact on the power system due to the instability and intermittence, and PV curtailment measures are taken to reduce the impact from voltage fluctuation. Accurate forecast is necessary to make a reasonable generation plan. A novel hybrid forecasting model is proposed, and an enhanced gray wolf optimization algorithm is proposed to improve the convergence ability and to solve the influence of extreme learning machine random parameters. The proposed algorithm has stronger convergence stability and higher convergence accuracy compared with the existing algorithms. The ensemble empirical mode decomposition algorithm is used to decompose the PV power under different weather conditions. The complexity of each component is calculated by the sample entropy, and the components are reconstructed to reduce the computational cost of forecasting models. The results revealed that mean absolute percentage error and root mean square error of the proposed model are smaller than 3% and 0.36 under various weather conditions. Meanwhile, the determination coefficient of the proposed model is more than 98%.

Assessing Nonlinear Dynamics and Trends in Precipitation by Ensemble Empirical Mode Decomposition (EEMD) and Fractal Approach in Benin Republic (West Africa)

Climate dynamics and trends have significant environmental and socioeconomic impacts; however, in the Benin Republic, they are generally studied with diverse statistical methods ignoring the nonstationarity, nonlinearity, and self-similarity characteristics contained in precipitation time series. This can lead to erroneous conclusions and an unclear understanding of climatic dynamics. Based on daily precipitation data observed in the six synoptic stations of Benin Republic, in the period from 1951 to 2010, we have proposed (i) determining the local trends of precipitations, (ii) investigating precipitation nonlinear dynamics, and (iii) assessing climatic shift in the study period by Ensemble Empirical Mode Decomposition (EEMD) and Multifractal Detrended Fluctuation Analysis (MFDFA) method. To overcome the detrending issue in the standard MFDFA method, the EEMD algorithm is embedded into the MFDFA. The study period is subdivided into three subperiods: 1951–1970, 1971–1990, and 1991–2010. Intrinsic Mode Functions (IMFs) are obtained according to the climatic region in which the stations are located. Results show that precipitation variation is significantly governed by the five first IMFs, in which oscillation periods vary from 1 to 25 days. The trend curves decrease at all the synoptic stations, and their slope values vary accordingly to the subperiods. Referring to the values of the multifractal spectrum parameters, α 0 , and the width of the spectrum w , consistent changes are observed regardless of the subperiods and the concerned stations. The spatial and temporal variability of precipitation indicates that the multifractal properties are good indicators for assessing changes in precipitation dynamics, and the changes in their features could be explained by the global change than by the local climate variation (climatic zones). Despite the observed differences in multifractal spectra properties from the three subperiods, it is not possible to verify the subdivision of 1951–2010 in three subperiods as it is done by previous studies in West Africa. Our findings can be used in the validation of global and regional climate models since a valid model should explain empirically detected scaling properties in observed data.

Research on the Denoising of the Beidou Carrier Signal based on EEMD Algorithm with Adaptive Reconstruction Optimizing of IMF Numbers

The Beidou carrier signal is coupled into a certain noise during propagation and reception, and these noise will directly affect the processing procedure associated with it. To deal with the problem of the influence due to the manually setting the IMF (Intrinsic Mode Function) component number for the reconstruction signal, a new measuring index that used for finding the optimal IMF components to reconstruct the signal has been designed in this paper. The index has taken the shape of the signal, signal noise ratio and correlation index into consideration. Upon on the basis, an adaptive index optimization Ensemble Empirical Mode Decomposition (AIO-EEMD) algorithm has been proposed in this paper. To verify the validity of the algorithm, four different algorithms are used to denoised the collected Beidou signal. The experiment results show that the noise reduction using the AIO-EEMD method can not only automatically obtain the optimal IMF components number, but also has a significant advantage over the other three methods.

Study on Denoising Algorithm for Power Quality Disturbances Based on Variational Mode Decomposition

The power quality signals are always mixed with noise, which will influence the accurate measurement and characteristic extraction. Therefore, it’s important to denoise for the signal processing. This paper puts forward a novel denoising algorithm for power quality disturbances (PQDs) based on variational mode decomposition (VMD) and the truth that the energy density times the average period of each intrinsic mode function (IMF) is a constant. Firstly, the VMD algorithm is used to decompose the white noise into several IMF components. Then, the average period energy product of each IMF component is calculated, and some of the IMF components are selected for reconstruction according to the judgment criterion to obtain the denoised signal. The simulation results show that the proposed algorithm can greatly improve the signal-to-noise ratio (SNR), and the denoising effect is better than that of wavelet threshold and ensemble empirical mode decomposition (EEMD) algorithm.

A fault frequency bands location method based on improved fast spectral correlation to extract fault features in axial piston pump bearings

In the working condition, the fault-excited impulses of axial piston pump bearings may be entirely buried by violent natural periodic impulses. Aiming at this problem, a fault frequency bands location method based on improved fast spectral correlation (Fast-SC) algorithm is proposed in this paper. Firstly, the Fast-SC is applied to analyze the original vibration signal to generate the cyclic spectral correlation image. Then, a new indicator named kurtosis enhanced spectral entropy (KESE) is exhibited to locate the fault frequency bands from the whole spectral frequency band, thereby highlighting the fault-excited impulses. Finally, the squared enhanced envelope spectrum (SEES) is employed to further extract feature frequencies and identify the fault type of the bearing. Experimental results validate the superiority and efficiency of the presented method. The fault-excited impulses extraction ability of the presented method is better than the traditional Fast-SC algorithm and ensemble empirical mode decomposition (EEMD) algorithm.

Switch-Based Low Intermediate Frequency System of a Vital Sign Radar for Simultaneous Multitarget and Multidirectional Detection

This paper proposes a radio frequency (RF) switch-based radar system for simultaneously detecting multitarget and multidirectional vital signs. Compared with traditional continuous wave (CW) radars that encounter mixed and distorted phases due to multiple reflections, frequency modulation CW (FMCW) radars have a suitable range resolution and relatively simple system architecture. Moreover, FMCW radars have the inherent capabilities of conducting multiple target monitoring and providing distance information. Based on our proposed algorithm, the vital signs of two targets at different distances from the antenna can be detected simultaneously even closely separated within the resolution bin limitation by simply using the proposed algorithm upon an FMCW radar to extract the vital signs separately. Moreover, instead of using multiple radar systems, a single transceiver can detect the vital signs from multiple directions and multiple targets by adding an RF switch as a multiplexer. This RF switch provides the mixing function. Therefore, a low intermediate frequency (IF) architecture was proposed by using the proposed RF switch-based radar system. This architecture simultaneously conducts multidirectional vital sign detection; thus, the signal-to-noise ratio (SNR) can be improved by appropriately selecting the IF value to reduce the flicker noise. The heart rate (HR) demodulation in a baseband is also challenging due to the low SNR and the respiratory rate (RR) harmonic for FMCW vital sign radars. Therefore, we applied an ensemble empirical mode decomposition algorithm to extract the intrinsic mode functions of RR and HR. Experimental results revealed that the proposed algorithm can significantly improve the SNR and accuracy. The vital sign errors of detecting multiple targets are less than 3% on average.

Stock price prediction using deep learning and frequency decomposition

Nonlinearity and high volatility of financial time series have made it difficult to predict stock price. However, thanks to recent developments in deep learning and methods such as long short-term memory (LSTM) and convolutional neural network (CNN) models, significant improvements have been obtained in the analysis of this type of data. Further, empirical mode decomposition (EMD) and complete ensemble empirical mode decomposition (CEEMD) algorithms decomposing time series to different frequency spectra are among the methods that could be effective in analyzing financial time series. Based on these theoretical frameworks, we propose novel hybrid algorithms, i.e., CEEMD-CNN-LSTM and EMD-CNN-LSTM, which could extract deep features and time sequences, which are finally applied to one-step-ahead prediction. The concept of the suggested algorithm is that when combining these models, some collaboration is established between them that could enhance the analytical power of the model. The practical findings confirm this claim and indicate that CNN alongside LSTM and CEEMD or EMD could enhance the prediction accuracy and outperform other counterparts. Further, the suggested algorithm with CEEMD provides better performance compared to EMD.

The baseline wander correction based on the improved ensemble empirical mode decomposition (EEMD) algorithm for grounded electrical source airborne transient electromagnetic signals

Abstract. The grounded electrical source airborne transient electromagnetic (GREATEM) system is an important method for obtaining subsurface conductivity distribution as well as outstanding detection efficiency and easy flight control. However, there are the superposition of desired signals and various noises for the GREATEM signal. The baseline wander caused by the receiving coil motion always exists in the process of data acquisition and affects measurement results. The baseline wander is one of the main noise sources, which has its own characteristics such as being low frequency, large amplitude, non-periodic, and non-stationary and so on. Consequently, it is important to correct the GREATEM signal for an inversion explanation. In this paper, we propose improving the method of ensemble empirical mode decomposition (EEMD) by adaptive filtering (EEMD-AF) based on EEMD to suppress baseline wander. Firstly, the EEMD-AF method will decompose the electromagnetic signal into multi-stage intrinsic mode function (IMF) components. Subsequently, the adaptive filter will process higher-index IMF components containing the baseline wander. Lastly, the de-noised signal will be reconstructed. To examine the performance of our introduced method, we processed the simulated and field signal containing the baseline wander by different methods. Through the evaluation of the signal-to-noise ratio (SNR) and mean-square error (MSE), the result indicates that the signal using the EEMD-AF method can get a higher SNR and lower MSE. Comparing correctional data using the EEMD-AF and the wavelet-based method in the anomaly curve profile images of the response signal, it is proved that the EEMD-AF method is practical and effective for the suppression of the baseline wander in the GREATEM signal.