Complete Ensemble Empirical Mode Decomposition With Adaptive Noise Research Articles

A long sequence time-series forecasting model for ship motion attitude based on informer

Influenced by marine environmental factors, a ship sailing at sea will experience six degrees of freedom oscillating motions. These oscillating motions pose safety risks to ship navigation and maritime operations. Especially in severe weather conditions, these oscillating motions have a significant impact on the takeoff and landing of carrier aircraft, as well as the launch of carrier missiles. Therefore, studying and forecasting the ship motion attitude in wind and waves holds immense significance. This paper presents a sparrow search algorithm (SSA) optimized Informer model based on complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) to forecast the long sequence time-series of ship motion attitude. Firstly, the ship motion attitude time series, as measured by the inertial navigation system, is decomposed using CEEMDAN to obtain several intrinsic mode functions (IMFs) and a residual. Then, the IMFs and residual are fed into SSA-Informer, where SSA is used to optimize the hyperparameter of Informer. Finally, the forecast values of IMFs and residual are superimposed and reconstructed to get the ship motion attitude forecasting result. The forecasting experiment under slight, moderate and rough sea conditions in the East China Sea shows that the proposed model achieves better performance and longer forecasting time compared with other popular methods.

Estimation of realized volatility of cryptocurrencies using CEEMDAN-RF-LSTM

Predicting cryptocurrency volatility is crucial for investors, traders, and decision-makers but is complicated by the market’s high non-linearity, volatility, and noise. This paper presents a novel approach, the CEEMDAN-RF-LSTM hybrid model, which is the first to combine the strengths of Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN), Random Forest (RF), and Long Short-Term Memory Network (LSTM) to predict the Realized Volatility (RV) of mainstream cryptocurrencies. The model exploits CEEMDAN’s proficiency in processing non-linear and non-stationary signals, RF’s exceptional feature selection capabilities, and LSTM’s distinctive advantages in dealing with time-series problems. Applied to actual transaction data for Bitcoin (BTC), Ethereum (ETH), and Binance Coin (BNB), empirical results show the superior performance of our model in predicting actual cryptocurrency volatility. These findings contribute to the academic understanding of cryptocurrency volatility and provide practical guidance for quantitative trading strategy development, offering fresh insights and methodologies for related research fields.

Comovement of african stock markets: Any influence from the COVID-19 pandemic?

Utilising daily data from twelve Sub-Saharan stock markets we investigate the co-movements and information transmission among African stock markets as a result of the impact of COVID while employing multiple wavelet techniques and applying the Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) to Renyi's and Shannon's effective transfer entropy analysis. The results infer that some number of co-movements exist among stock markets in Africa and that during periods of uncertainties, diversification through the creation of portfolios in African markets is not conducive since they tend to comove strongly during such periods. The study discovered that, a few of the markets responded to the pandemic in leads lags in the pre-, during and post-COVID era, as well as reacted to information transmission. Our findings generally show that information transmission/spillovers are more predominant in the short term than in the medium- and long-term horizons. The Renyi's effective transfer entropy recorded more negative information flows between African stock market than positive information flows, both during the COVID period and after. On the other hand, Shannon's entropy showed non-negative information flow across various time horizons. We conclude that even though most African stock markets were not prone to the contagion effect of the pandemic, it is of vital importance to re-evaluate the notion that African stock markets are immune to contagion of stock market co-movements, especially in times of global uncertainties.

A self-attention-LSTM method for dam deformation prediction based on CEEMDAN optimization

The structural deformation of a dam directly affects its lifespan and safety, making its accurate prediction crucial. Traditional prediction methods often overlook the nonlinearity and non-smoothness of deformation data. Moreover, the irregular intervals within the historical deformation data used for model training can reduce prediction accuracy. To address these issues, we propose a hybrid deep learning model that uses signal decomposition and reconstruction to enhance dam deformation prediction. This model employs a long short-term memory (LSTM) neural network optimized using a complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN). The self-attention mechanism in the LSTM model effectively captures the temporal features of dam deformation, alleviating the difficulties generated by the irregular intervals within the historical data used in model training. Furthermore, considering the lag effect of influencing factors on dam deformation and the differences among various measurement points, we propose a CEEMDAN-based feature selection method. Using 13 years worth of data from the Shuibuya Dam, we evaluate the accuracy and effectiveness of the CEEMDAN-Self-attention-LSTM model using indicators, such as MAE, RMSE, MAPE, and R2, and compared it with existing models. The experimental results show that this model reduces prediction error by more than 53.62%.

An Ultra-Short-Term Wind Power Prediction Method Based on Quadratic Decomposition and Multi-Objective Optimization

To augment the accuracy, stability, and qualification rate of wind power prediction, thereby fostering the secure and economical operation of wind farms, a method predicated on quadratic decomposition and multi-objective optimization for ultra-short-term wind power prediction is proposed. Initially, the original wind power signal is decomposed using a quadratic decomposition method constituted by the Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN), Fuzzy Entropy (FE), and Symplectic Geometry Mode Decomposition (SGMD), thereby mitigating the randomness and volatility of the original signal. Subsequently, the decomposed signal components are introduced into the Deep Bidirectional Long Short-Term Memory (DBiLSTM) neural network for time series modeling, and the Sand Cat Swarm Optimization Algorithm (SCSO) is employed to optimize the network hyperparameters, thereby enhancing the network’s predictive performance. Ultimately, a multi-objective optimization loss that accommodates accuracy, stability, and grid compliance is proposed to guide network training. Experimental results reveal that the employed quadratic decomposition method and the proposed multi-objective optimization loss can effectively bolster the model’s predictive performance. Compared to other classical methods, the proposed method achieves optimal results across different seasons, thereby demonstrating robust practicality.

Research on Speech Denoising Algorithm for Equipment Trains in Fully Mechanized Mining Working Faces based on CEEMDAN High and Low Frequency Filtering

This paper proposes a high and low frequency filtering denoising algorithm based on Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN). CEEMDAN to address the issue of noise in speech signals obtained in the environment of fully mechanized mining face equipment trains. This algorithm first decomposes speech signals into multiple Intrinsic Mode Functions (IMFs) in different frequency domains through CEEMDAN, then uses Hilbert-Huang Transform (HHT) to obtain the instantaneous frequency and amplitude of IMFs, calculates the Euclidean distance between IMFs, normalizes it to obtain the average distance between IMFs, and divides the IMFs components into high and low frequency categories based on the average distance. Next, Optimally Modified Log Spectral Amplitude (OMLSA) algorithm and Kalman filtering algorithm are used for the high-frequency and low-frequency IMF components, respectively. Finally, the processed IMF components are reconstructed to obtain a denoised signal. After comparative experiments, the noise reduction algorithm proposed in this paper significantly improves the Signal to Noise Ratio (SNR), Peak Signal to Noise Ratio (PSNR), Mean Square Error (RMSE), and Mean Absolute Distance (MAE) of speech signals obtained in the environment of fully mechanized mining face equipment trains compared to other commonly used noise reduction algorithms.

Short-Term Electric Load Forecasting Based on Signal Decomposition and Improved TCN Algorithm

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.

An Improved Identification Method of Pipeline Leak Using Acoustic Emission Signal

Pipelines constitute a vital component in offshore oil and gas operations, subjected to prolonged exposure to a range of alternating loads. Safeguarding their integrity, particularly through meticulous leak detection, is essential for ensuring safe and reliable operation. Acoustic emission detection emerges as an effective approach for monitoring pipeline leaks, demanding subsequent rigorous data analysis. Traditional analysis techniques like wavelet analysis, empirical mode decomposition (EMD), variational mode decomposition (VMD), and complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) often yield results with considerable randomness, adversely affecting leak detection accuracy. This study introduces an enhanced damage recognition methodology, integrating improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) and probabilistic neural networks (PNN) for more accurate pipeline leak identification. This novel approach combines laboratory-acquired acoustic emission signals from leaks with ambient noise signals. Application of ICEEMDAN to these composite signals isolates eight intrinsic mode functions (IMFs), with subsequent time–frequency analysis providing insight into their frequency structures and feature vectors. These vectors are then employed to train a PNN, culminating in a robust neural network model tailored for leak detection. Conduct experimental research on pipeline leakage identification, focusing on the local structure of offshore platforms, experimental research validates the superiority of the ICEEMDAN–PNN model over existing methods like EMD, VMD, and CEEMDAN paired with PNN, particularly in terms of stability, anti-interference capabilities, and detection precision. Notably, even amidst integrated noise, the ICEEMDAN–PNN model maintains a remarkable 98% accuracy rate in identifying pipeline leaks.

Short-term load forecasting based on CEEMDAN and dendritic deep learning

Short-term load forecasting (STLF) plays a key role in improving the operational efficiency of real-time electricity markets by reducing load uncertainty, optimising the allocation of generation resources and increasing the reliability of electricity supply. Despite significant advances in the field, achieving high accuracy in STLF remains a challenge due to the inherent complexity and volatility of load data. This study proposes a novel STLF method that synergistically combines the strengths of a dendritic neuron model (DNM), long short-term memory, and complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) to address these challenges. First, historical load data is decomposed into components of different frequencies using CEEMDAN. Then, LSTM is used to extract the temporal characteristics of these load subsequence components in combination with various influencing factors such as temperature and humidity. Uniquely, this study innovatively integrates a dendritic neuron model instead of the traditional fully connected layer, a structural innovation that enhances the model’s ability to explore the intrinsic correlations of the time series data in depth. This allows the model to process and predict each component individually, and then later reconstruct and sum these independent prediction results to obtain a final prediction value with higher accuracy. Experiments are conducted on the Panama electricity load dataset to compare the proposed dendritic learning model with several state-of-the-art models, including CrossFormer and Transformer. The results confirm that the proposed model has significantly better prediction accuracy.

A DENOISING METHOD FOR COLONIC PRESSURE SIGNAL BASED ON CEEMDAN-WT

Colonic pressure is a principal parameter reflecting intestinal motility, and is also an important basis for diagnosing intestinal motility disorders. The physiological signals of the colon are mixed with the disturbing signals such as breathing and heartbeat due to the complex environment of the intestine. It is essential to find an effective method to reduce the noise in the signals. To solve this problem, a combined denoising method of colonic pressure signals is proposed in this paper, which combines the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and wavelet transform. In this method, colonic peristalsis pressure signal is first separated out by wavelet transform, then decomposed by CEEMDAN. According to pearson correlation coefficient, the high-frequency intrinsic mode functions (IMFs) with noise was selected and denoised by wavelet threshold. Finally, the signal was reconstructed by combining the reserved IMFs with the denoised IMFs. The validity of the proposed method is verified by the real colon pressure data collected by the wireless motility capsule in this paper. The experimental results show that the combined denoising method can deal with the noise better than the single denoising method on the premise of fully retaining the real components of the colonic pressure signal.

Enhanced complete ensemble EMD with superior noise handling capabilities: A robust signal decomposition method for power systems analysis

AbstractSignal decomposition is crucial in several domains, particularly in the dissection of complex signals present in electrical power systems. Understanding the oscillations and patterns within these signals can significantly influence energy resource management, grid stability, and efficient system operation. This paper presents an advanced enhanced decomposition method based on Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) to mitigate the inherent drawbacks of the conventional CEEMDAN and its improved version. Unlike CEEMDAN's generalized noise approach, the proposed method introduces adaptive noise, enhancing target signal noise handling by incorporating a tailored filtering and updating process after each iteration. This leads to more accurate signal decomposition compared to traditional methods. Comprehensive tests were conducted using artificially generated signals characterized by mode mixing, varying frequency oscillations, complex real‐world electrical demand signals, generator axis vibrations and partial discharge signals. The results demonstrate that the proposed method outperforms traditional techniques in two significant aspects. First, it provides superior spectral separation of the intrinsic modes (IMF) of the signal, thereby enhancing decomposition accuracy. Second, it significantly reduced the number of shifting iterations, thereby alleviating the computational load. These advancements have led to a more accurate and efficient framework that is essential for analyzing nonlinear and nonstationary signals.

AIoT-driven multi-source sensor emission monitoring and forecasting using multi-source sensor integration with reduced noise series decomposition

The integration of multi-source sensors based AIoT (Artificial Intelligence of Things) technologies into air quality measurement and forecasting is becoming increasingly critical in the fields of sustainable and smart environmental design, urban development, and pollution control. This study focuses on enhancing the prediction of emission, with a special emphasis on pollutants, utilizing advanced deep learning (DL) techniques. Recurrent neural networks (RNNs) and long short-term memory (LSTM) neural networks have shown promise in predicting air quality trends in time series data. However, challenges persist due to the unpredictability of air quality data and the scarcity of long-term historical data for training. To address these challenges, this study introduces the AIoT-enhanced EEMD-CEEMDAN-GCN model. This innovative approach involves decomposing the input signal using EEMD (Ensemble Empirical Mode Decomposition) and CEEMDAN (Complete Ensemble Empirical Mode Decomposition with Adaptive Noise) to extract intrinsic mode functions. These functions are then processed through a GCN (Graph Convolutional Network) model, enabling precise prediction of air quality trends. The model’s effectiveness is validated using air pollution datasets from four provinces in China, demonstrating its superiority over various deep learning models (GCN, EMD-GCN) and series decomposition models (EEMD-GCN, CEEMDAN-GCN). It achieves higher accuracy and better data fitting, outperforming other models in key metrics such as MAE (Mean Absolute Error), MSE (Mean Squared Error), MAPE (Mean Absolute Percentage Error), and R2 (Coefficient of Determination). The implementation of this AIoT-enhanced model in air pollution prediction allows decision-makers to more accurately anticipate changes in air quality, particularly concerning carbon emissions. This facilitates more effective planning of mitigation measures, improvement of public health, and optimization of resource allocation. Moreover, the model adeptly addresses the complexities of air quality data, contributing significantly to enhanced monitoring and management strategies in the context of sustainable urban development and environmental conservation.

A hybrid short-term load forecasting method using CEEMDAN-RCMSE and improved BiLSTM error correction

Accurate load forecasting is an important issue for safe and economic operation of power system. However, load data often has strong non-stationarity, nonlinearity and randomness, which increases the difficulty of load forecasting. To improve the prediction accuracy, a hybrid short-term load forecasting method using load feature extraction based on complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and refined composite multi-scale entropy (RCMSE) and improved bidirectional long short time memory (BiLSTM) error correction is proposed. Firstly, CEEMDAN is used to separate the detailed information and trend information of the original load series, RCMSE is used to reconstruct the feature information, and Spearman is used to screen the features. Secondly, an improved butterfly optimization algorithm (IBOA) is proposed to optimize BiLSTM, and the reconstructed components are predicted respectively. Finally, an error correction model is constructed to mine the hidden information contained in error sequence. The experimental results show that the MAE, MAPE and RMSE of the proposed method are 645 kW, 0.96% and 827.3 kW respectively, and MAPE is improved by about 10% compared with other hybrid models. Therefore, the proposed method can overcome the problem of inaccurate prediction caused by data and inherent defects of models and improve the prediction accuracy.

Forecasting price in a new hybrid neural network model with machine learning

A key aspect of asset investment and risk management is the study of forecasting stock prices. We investigate the machine learning stock price prediction in a new hybrid neural network model and put forth a forecasting method based on machine learning, composite data preprocessing method and the proposed new neural network model. To address the challenge of predicting stock prices in the face of market complexity and noise, we use the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) algorithm and the Savitzky–Golay (SG) filter to de-noise and enhance the data, and employ a neural network with three convolutional layers and a long short-term memory (LSTM) layer that enables it to capture complex temporal patterns in the data. We propose a new hybrid neural network prediction model (CEEMDAN-S-C-LSTM) and adopt a machine learning approach to compare it with the benchmark model using CSI 300 index data. The empirical results validate the effectiveness of the frequency decomposition algorithm and the convolutional layer, and demonstrate that our proposed model outperforms the benchmark model. Compared to the best benchmark model, the CEEMDAN-S-C-LSTM model proposed in this study demonstrates a significant improvement in performance. Specifically, it shows a 45.33% reduction in mean absolute error (MAE), 43.44% reduction in root mean square error (RMSE), 45.01% reduction in mean absolute percentage error (MAPE), and 3.90% improvement in coefficient of determination (R2). The study also explores the effect of different numbers of convolution layers and SG filters on the hybrid model. Our research expands on the use of neural networks and machine learning, offering a novel technical approach to making investment decisions and managing risks in financial systems.

Short-term load forecasting method based on secondary decomposition and improved hierarchical clustering

In the context of large-scale grid connection of new energy, short-term load forecasting is a vital and challenging task for power system to balance supply and demand. To effectively improve the forecasting accuracy, a new load forecasting method is proposed aiming to mine the characteristics of load data and study the application of artificial intelligence algorithms. In this paper, the seasonal and trend decomposition using loess (STL) method is firstly applied to decompose the load data into the trend, seasonal and residual components and the residual component with the highest complexity is further decomposed by the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) approach. Secondly, in order to reduce the number of components, the improved hierarchical clustering technique is proposed to cluster all intrinsic mode functions (IMFs) obtained by CEEMDAN into high-frequency and low-frequency components. Then, different network models are trained to get the prediction results for different components, and the total load prediction value is achieved by stacking all of them. Finally, the national demand dataset of Great Britain in 2021–2022 is used to conduct the ablation and comparative experiments. The mean absolute percentage error (MAPE) and the root mean square error (RMSE) of the proposed method are 2.064% and 724.01 MW, respectively, which verified the effectiveness and advancement of the proposed method.

A Novel Ensemble Machine Learning Model for Oil Production Prediction with Two-Stage Data Preprocessing

Petroleum production forecasting involves the anticipation of fluid production from wells based on historical data. Compared to traditional empirical, statistical, or reservoir simulation-based models, machine learning techniques leverage inherent relationships among historical dynamic data to predict future production. These methods are characterized by readily available parameters, fast computational speeds, high precision, and time–cost advantages, making them widely applicable in oilfield production. In this study, time series forecast models utilizing robust and efficient machine learning techniques are formulated for the prediction of production. We have fused the two-stage data preprocessing methods and the attention mechanism into the temporal convolutional network-gated recurrent unit (TCN-GRU) model. Firstly, the random forest (RF) algorithm is employed to extract key dynamic production features that influence output, serving to reduce data dimensionality and mitigate overfitting. Next, the mode decomposition algorithm, complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN), is introduced. It employs a decomposition–reconstruction approach to segment production data into high-frequency noise components, low-frequency regular components and trend components. These segments are then individually subjected to prediction tasks, facilitating the model’s ability to capture more accurate intrinsic relationships among the data. Finally, the TCN-GRU-MA model, which integrates a multi-head attention (MA) mechanism, is utilized for production forecasting. In this model, the TCN module is employed to capture temporal data features, while the attention mechanism assigns varying weights to highlight the most critical influencing factors. The experimental results indicate that the proposed model achieves outstanding predictive performance. Compared to the best-performing comparative model, it exhibits a reduction in RMSE by 3%, MAE by 1.6%, MAPE by 12.7%, and an increase in R2 by 2.6% in Case 1. Similarly, in Case 2, there is a 7.7% decrease in RMSE, 7.7% in MAE, 11.6% in MAPE, and a 4.7% improvement in R2.

A hybrid prediction model of dissolved oxygen concentration based on secondary decomposition and bidirectional gate recurrent unit.

Dissolved oxygen is one of the important comprehensive indicators of river water quality, which reflects the degree of pollution in the water body. Monitoring and predicting dissolved oxygen are an important tool for water quality management, which helps to effectively maintain water ecological balance and prevent environmental problems. A single model cannot describe the dynamic characteristics of dissolved oxygen sequence, which affects the prediction accuracy. In order to obtain more accurate dissolved oxygen prediction results, decomposition techniques are commonly used to extract the main fluctuations and trends of water quality sequences. However, the high-frequency modes obtained from decomposition are still unstable. To solve this problem, this paper proposed a hybrid prediction model of dissolved oxygen concentration based on secondary decomposition and bidirectional gate recurrent unit. Firstly, dissolved oxygen sequence is preliminarily decomposed by complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and obtain several intrinsic mode functions (IMF). The fuzzy entropy (FE) is calculated to quantify the complexity of the IMF. Then, variational mode decomposition improved by northern goshawk optimization is used to decompose the IMF with higher entropy. The nonlinearity and instability of the sequence are further weakened. Finally, the bidirectional gate recurrent unit (BiGRU) neural network is used to predict each IMF component, and the final prediction result is obtained by reconstructing the prediction results of each component. In order to verify the effectiveness of the proposed model, this paper selects the dissolved oxygen data of Xin'anjiang Reservoir as the research object. The experimental results show that the RMSE, MAE, MAPE, and R2 of the proposed model are 0.1164, 0.0894, 1.0403%, and 0.9939, respectively, which is best among other comparative prediction models (BP, LSTM, GRU, BiGRU, EMD-BiGRU, CEEMDAN-BiGRU, VMD-BiGRU, and GNO-VMD-BiGRU). Therefore, this model effectively deals with high volatility and nonlinear dissolved oxygen data and provides reference for water environment management and ecological protection.

Permutation fuzzy entropy based ICEEMDAN de-noising for inertial sensors

As one of the critical factors affecting the accuracy of inertial navigation system (INS) positioning, random noise will lead to the accumulation of INS positioning errors, even the failure of the system. Previous studies mainly focus on achieving noise reduction by adopting either the wavelet decomposition method or some relevant methods of the empirical mode decomposition (EMD). However, the former is limited by Heisenberg’s uncertainty principle when processing non-stationary signals, and the latter exhibits numerous mode-mixing phenomena during the decomposition process. In this paper, a permutation fuzzy entropy (PFE) based improved complete ensemble EMD with adaptive noise (ICEEMDAN) de-noising method is proposed to utilize the ICEEMDAN method for signal decomposition, and the PFE as an index to distinguish the effective intrinsic mode functions (IMFs) and the noisy IMFs during the decomposition process. In addition, simulated signals with different noise levels and field data with different durations are utilized to design a group of experiments to test the de-noising performance of the PFE-based ICEEMDAN de-noising method. The experimental results demonstrate that the advantages of our proposed method consists of three main parts: (1) The ICEEMDAN method more effectively alleviates the mode-mixing problem compared to the complete ensemble EMD with adaptive noise (CEEMDAN), thereby enhancing the accuracy of inertial measurement unit (IMU) signal reconstruction; (2) the PFE is more preferable than the permutation entropy for classify IMFs decomposed by the ICEEMDAN method; (3) our proposed method can effectively reduce IMU signal noise and improve the positioning accuracy of inertial sensor.

Prediction model of land surface settlement deformation based on improved LSTM method: CEEMDAN-ICA-AM-LSTM (CIAL) prediction model.

The uneven settlement of the surrounding ground surface caused by subway construction is not only complicated but also liable to cause casualties and property damage, so a timely understanding of the ground settlement deformation in the subway excavation and its prediction in real time is of practical significance. Due to the complex nonlinear relationship between subway settlement deformation and numerous influencing factors, as well as the existence of a time lag effect and the influence of various factors in the process, the prediction performance and accuracy of traditional prediction methods can no longer meet industry demands. Therefore, this paper proposes a surface settlement deformation prediction model by combining noise reduction and attention mechanism (AM) with the long short-term memory (LSTM). The complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and independent component analysis (ICA) methods are used to denoise the input original data and then combined with AM and LSTM for prediction to obtain the CEEMDAN-ICA-AM-LSTM (CIAL) prediction model. Taking the settlement monitoring data of the construction site of Urumqi Rail Transit Line 1 as an example for analysis reveals that the model in this paper has better effectiveness and applicability in the prediction of surface settlement deformation than multiple prediction models. The RMSE, MAE, and MAPE values of the CIAL model are 0.041, 0.033 and 0.384%; R2 is the largest; the prediction effect is the best; the prediction accuracy is the highest; and its reliability is good. The new method is effective for monitoring the safety of surface settlement deformation.

Short-Term Energy Consumption Prediction of Large Public Buildings Combined with Data Feature Engineering and Bilstm-Attention

Accurate building energy consumption prediction is a crucial condition for the sustainable development of building energy management systems. However, the highly nonlinear nature of data and complex influencing factors in the energy consumption of large public buildings often pose challenges in improving prediction accuracy. In this study, we propose a combined prediction model that combines signal decomposition, feature screening, and deep learning. First, we employ the Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) to decompose energy consumption data. Next, we propose the Maximum Mutual Information Coefficient (MIC)-Fast Correlation Based Filter (FCBF) combined feature screening method for feature selection on the decomposed components. Finally, the selected input features and corresponding components are fed into the Bi-directional Long Short-Term Memory Attention Mechanism (BiLSTMAM) model for prediction, and the aggregated results yield the energy consumption forecast. The proposed approach is validated using energy consumption data from a large public building in Shaanxi Province, China. Compared with the other five comparison methods, the RMSE reduction of the CEEMDAN-MIC-FCBF-BiLSTMAM model proposed in this study ranged from 57.23% to 82.49%. Experimental results demonstrate that the combination of CEEMDAN, MIC-FCBF, and BiLSTMAM modeling markedly improves the accuracy of energy consumption predictions in buildings, offering a potent method for optimizing energy management and promoting sustainability in large-scale facilities.