Complicated loads encountered by floating offshore wind turbines(FOWTs)in real sea conditions are crucial for future optimization of design,but obtaining data on them directly poses a challenge.To address this issue,w...Complicated loads encountered by floating offshore wind turbines(FOWTs)in real sea conditions are crucial for future optimization of design,but obtaining data on them directly poses a challenge.To address this issue,we applied machine learning techniques to obtain hydrodynamic and aerodynamic loads of FOWTs by measuring platform motion responses and wave-elevation sequences.First,a computational fluid dynamics(CFD)simulation model of the floating platform was established based on the dynamic fluid body interaction technique and overset grid technology.Then,a long short-term memory(LSTM)neural network model was constructed and trained to learn the nonlinear relationship between the waves,platform-motion inputs,and hydrodynamic-load outputs.The optimal model was determined after analyzing the sensitivity of parameters such as sample characteristics,network layers,and neuron numbers.Subsequently,the effectiveness of the hydrodynamic load model was validated under different simulation conditions,and the aerodynamic load calculation was completed based on the D'Alembert principle.Finally,we built a hybrid-scale FOWT model,based on the software in the loop strategy,in which the wind turbine was replaced by an actuation system.Model tests were carried out in a wave basin and the results demonstrated that the root mean square errors of the hydrodynamic and aerodynamic load measurements were 4.20%and 10.68%,respectively.展开更多
Intelligent maintenance of roads and highways requires accurate deterioration evaluation and performance prediction of asphalt pavement.To this end,we develop a time series long short-term memory(LSTM)model to predict...Intelligent maintenance of roads and highways requires accurate deterioration evaluation and performance prediction of asphalt pavement.To this end,we develop a time series long short-term memory(LSTM)model to predict key performance indicators(PIs)of pavement,namely the international roughness index(IRI)and rutting depth(RD).Subsequently,we propose a comprehensive performance indicator for the pavement quality index(PQI),which leverages the highway performance assessment standard method,entropy weight method,and fuzzy comprehensive evaluation method.This indicator can evaluate the overall performance condition of the pavement.The data used for the model development and analysis are extracted from tests on two full-scale accelerated test tracks,called MnRoad and RIOHTrack.Six variables are used as predictors,including temperature,precipitation,total traffic volume,asphalt surface layer thickness,pavement age,and maintenance condition.Furthermore,wavelet denoising is performed to analyze the impact of missing or abnormal data on the LSTM model accuracy.In comparison to a traditional autoregressive integrated moving average(ARIMAX)model,the proposed LSTM model performs better in terms of PI prediction and resiliency to noise.Finally,the overall prediction accuracy of our proposed performance indicator PQI is 93.8%.展开更多
Deep learning plays a vital role in real-life applications, for example object identification, human face recognition, speech recognition, biometrics identification, and short and long-term forecasting of data. The ma...Deep learning plays a vital role in real-life applications, for example object identification, human face recognition, speech recognition, biometrics identification, and short and long-term forecasting of data. The main objective of our work is to predict the market performance of the Dhaka Stock Exchange (DSE) on day closing price using different Deep Learning techniques. In this study, we have used the LSTM (Long Short-Term Memory) network to forecast the data of DSE for the convenience of shareholders. We have enforced LSTM networks to train data as well as forecast the future time series that has differentiated with test data. We have computed the Root Mean Square Error (RMSE) value to scrutinize the error between the forecasted value and test data that diminished the error by updating the LSTM networks. As a consequence of the renovation of the network, the LSTM network provides tremendous performance which outperformed the existing works to predict stock market prices.展开更多
This research focuses on solving the fault detection and health monitoring of high-power thyristor converter.In terms of the critical role of thyristor converter in nuclear fusion system,a method based on long short-t...This research focuses on solving the fault detection and health monitoring of high-power thyristor converter.In terms of the critical role of thyristor converter in nuclear fusion system,a method based on long short-term memory(LSTM)neural network model is proposed to monitor the operational state of the converter and accurately detect faults as they occur.By sampling and processing a large number of thyristor converter operation data,the LSTM model is trained to identify and detect abnormal state,and the power supply health status is monitored.Compared with traditional methods,LSTM model shows higher accuracy and abnormal state detection ability.The experimental results show that this method can effectively improve the reliability and safety of the thyristor converter,and provide a strong guarantee for the stable operation of the nuclear fusion reactor.展开更多
The problems in equipment fault detection include data dimension explosion,computational complexity,low detection accuracy,etc.To solve these problems,a device anomaly detection algorithm based on enhanced long short-...The problems in equipment fault detection include data dimension explosion,computational complexity,low detection accuracy,etc.To solve these problems,a device anomaly detection algorithm based on enhanced long short-term memory(LSTM)is proposed.The algorithm first reduces the dimensionality of the device sensor data by principal component analysis(PCA),extracts the strongly correlated variable data among the multidimensional sensor data with the lowest possible information loss,and then uses the enhanced stacked LSTM to predict the extracted temporal data,thus improving the accuracy of anomaly detection.To improve the efficiency of the anomaly detection,a genetic algorithm(GA)is used to adjust the magnitude of the enhancements made by the LSTM model.The validation of the actual data from the pumps shows that the algorithm has significantly improved the recall rate and the detection speed of device anomaly detection,with the recall rate of 97.07%,which indicates that the algorithm is effective and efficient for device anomaly detection in the actual production environment.展开更多
This paper introduces the time-frequency analyzed long short-term memory(TF-LSTM) neural network method for jamming signal recognition over the Global Navigation Satellite System(GNSS) receiver. The method introduces ...This paper introduces the time-frequency analyzed long short-term memory(TF-LSTM) neural network method for jamming signal recognition over the Global Navigation Satellite System(GNSS) receiver. The method introduces the long shortterm memory(LSTM) neural network into the recognition algorithm and combines the time-frequency(TF) analysis for signal preprocessing. Five kinds of navigation jamming signals including white Gaussian noise(WGN), pulse jamming, sweep jamming, audio jamming, and spread spectrum jamming are used as input for training and recognition. Since the signal parameters and quantity are unknown in the actual scenario, this work builds a data set containing multiple kinds and parameters jamming to train the TF-LSTM. The performance of this method is evaluated by simulations and experiments. The method has higher recognition accuracy and better robustness than the existing methods, such as LSTM and the convolutional neural network(CNN).展开更多
Load forecasting is of great significance to the development of new power systems.With the advancement of smart grids,the integration and distribution of distributed renewable energy sources and power electronics devi...Load forecasting is of great significance to the development of new power systems.With the advancement of smart grids,the integration and distribution of distributed renewable energy sources and power electronics devices have made power load data increasingly complex and volatile.This places higher demands on the prediction and analysis of power loads.In order to improve the prediction accuracy of short-term power load,a CNN-BiLSTMTPA short-term power prediction model based on the Improved Whale Optimization Algorithm(IWOA)with mixed strategies was proposed.Firstly,the model combined the Convolutional Neural Network(CNN)with the Bidirectional Long Short-Term Memory Network(BiLSTM)to fully extract the spatio-temporal characteristics of the load data itself.Then,the Temporal Pattern Attention(TPA)mechanism was introduced into the CNN-BiLSTM model to automatically assign corresponding weights to the hidden states of the BiLSTM.This allowed the model to differentiate the importance of load sequences at different time intervals.At the same time,in order to solve the problem of the difficulties of selecting the parameters of the temporal model,and the poor global search ability of the whale algorithm,which is easy to fall into the local optimization,the whale algorithm(IWOA)was optimized by using the hybrid strategy of Tent chaos mapping and Levy flight strategy,so as to better search the parameters of the model.In this experiment,the real load data of a region in Zhejiang was taken as an example to analyze,and the prediction accuracy(R2)of the proposed method reached 98.83%.Compared with the prediction models such as BP,WOA-CNN-BiLSTM,SSA-CNN-BiLSTM,CNN-BiGRU-Attention,etc.,the experimental results showed that the model proposed in this study has a higher prediction accuracy.展开更多
To fully make use of information from different representation subspaces,a multi-head attention-based long short-term memory(LSTM)model is proposed in this study for speech emotion recognition(SER).The proposed model ...To fully make use of information from different representation subspaces,a multi-head attention-based long short-term memory(LSTM)model is proposed in this study for speech emotion recognition(SER).The proposed model uses frame-level features and takes the temporal information of emotion speech as the input of the LSTM layer.Here,a multi-head time-dimension attention(MHTA)layer was employed to linearly project the output of the LSTM layer into different subspaces for the reduced-dimension context vectors.To provide relative vital information from other dimensions,the output of MHTA,the output of feature-dimension attention,and the last time-step output of LSTM were utilized to form multiple context vectors as the input of the fully connected layer.To improve the performance of multiple vectors,feature-dimension attention was employed for the all-time output of the first LSTM layer.The proposed model was evaluated on the eNTERFACE and GEMEP corpora,respectively.The results indicate that the proposed model outperforms LSTM by 14.6%and 10.5%for eNTERFACE and GEMEP,respectively,proving the effectiveness of the proposed model in SER tasks.展开更多
Load forecasting can increase the efficiency of modern energy systems with built-in measuring systerms by providing a more accurate peak power shaving performance and thus more reliable control.An analysis of an integ...Load forecasting can increase the efficiency of modern energy systems with built-in measuring systerms by providing a more accurate peak power shaving performance and thus more reliable control.An analysis of an integrated CO2 heat pump and chiller system with a hot water storage system is presented in this paper.Drastic power fluctuations,which can be reduced with load forecasting,are found in historical operation records.A model that aims to forecast the ventilation system heating demand is thus established on the basis of a long short-term memory(LSTM)network.The model can successfully forecast the one hour ahead power using records of the past 48h of the system operation data and the ambient temperature.The mean absolute percentage error(MAPE)of the forecast results of the LSTM-based model is 10.70%,which is respectively 2.2%and 7.25%better than the MAPEs of the forecast results of the support vector regression based and persistence method based models.展开更多
Power prediction has been critical in large-scale wind power grid connections.However,traditional wind power prediction methods have long suffered from problems,for instance low prediction accuracy and poor reliabilit...Power prediction has been critical in large-scale wind power grid connections.However,traditional wind power prediction methods have long suffered from problems,for instance low prediction accuracy and poor reliability.For this purpose,a hybrid prediction model(VMD-LSTM-Attention)has been proposed,which integrates the variational modal decomposition(VMD),the long short-term memory(LSTM),and the attention mechanism(Attention),and has been optimized by improved dung beetle optimization algorithm(IDBO).Firstly,the algorithm's performance has been significantly enhanced through the implementation of three key strategies,namely the elite group strategy of the Logistic-Tent map,the nonlinear adjustment factor,and the adaptive T-distribution disturbance mechanism.Subsequently,IDBO has been applied to optimize the important parameters of VMD(decomposition layers and penalty factors)to ensure the best decomposition signal is obtained;Furthermore,the IDBO has been deployed to optimize the three key hyper-parameters of the LSTM,thereby improving its learning capability.Finally,an Attention mechanism has been incorporated to adaptively weight temporal features,thus increasing the model's ability to focus on key information.Comprehensive simulation experiments have demonstrated that the proposed model achieves higher prediction accuracy compared with VMD-LSTM,VMD-LSTM-Attention,and traditional prediction methods,and quantitative indexes verify the efectiveness of the algorithmic improvement as well as the excellence and precision of the model in wind power prediction.展开更多
Stroke survivors often face significant challenges when performing daily self-care activities due to upper limb motor impairments.Traditional surface electromyography(sEMG)analysis typically focuses on isolated hand p...Stroke survivors often face significant challenges when performing daily self-care activities due to upper limb motor impairments.Traditional surface electromyography(sEMG)analysis typically focuses on isolated hand postures,overlooking the complexity of object-interactive behaviors that are crucial for promoting patient independence.This study introduces a novel framework that combines high-density sEMG(HD-sEMG)signals with an improved Whale Optimization Algorithm(IWOA)-optimized Long Short-Term Memory(LSTM)network to address this limitation.The key contributions of this work include:(1)the creation of a specialized HD-sEMG dataset that captures nine continuous self-care behaviors,along with time and posture markers,to better reflect real-world patient interactions;(2)the development of a multi-channel feature fusion module based on Pascal’s theorem,which enables efficient signal segmentation and spatial–temporal feature extraction;and(3)the enhancement of the IWOA algorithm,which integrates optimal point set initialization,a diversity-driven pooling mechanism,and cosine-based differential evolution to optimize LSTM hyperparameters,thereby improving convergence and global search capabilities.Experimental results demonstrate superior performance,achieving 99.58%accuracy in self-care behavior recognition and 86.19%accuracy for 17 continuous gestures on the Ninapro db2 benchmark.The framework operates with low latency,meeting the real-time requirements for assistive devices.By enabling precise,context-aware recognition of daily activities,this work advances personalized rehabilitation technologies,empowering stroke patients to regain autonomy in self-care tasks.The proposed methodology offers a robust,scalable solution for clinical applications,bridging the gap between laboratory-based gesture recognition and practical,patient-centered care.展开更多
Deep learning enables neural networks to improve prediction performance through data supplementation.In financial time series forecasting,however,such data-driven approaches can encounter limitations where additional ...Deep learning enables neural networks to improve prediction performance through data supplementation.In financial time series forecasting,however,such data-driven approaches can encounter limitations where additional data degrade performance,contrary to common expectations.While more data can still be beneficial,it may introduce systemic concept drift due to the complex nonstationarities of stock price index time series,thereby exacerbating overfitting.One such drift is memory inconsistency:locally measured long memories fluctuate over time,alternately approaching and deviating from the random walk condition.We address this problem by typifying memory inconsistencies into two simplified forms:long-term dependentto-independent(D2I)and long-term independent-to-dependent(I2D)inconsistencies.The first experiment,which uses U.S.stock price indices,suggests that additional training examples may lead to performance deterioration of long short-term memory(LSTM)networks,especially when memory inconsistencies are prominent.Since stock markets are influenced by numerous unknown dynamics,the second experiment,which uses simulated mean-reverting time series derived from the fractional Ornstein–Uhlenbeck(fOU)process,is conducted to focus solely on challenges arising from memory inconsistencies.The experimental results demonstrate that memory inconsistencies disrupt the performance of LSTM networks.Theoretically,additional errors from D2I and I2D inconsistencies increase as the time lag increases.Since LSTM networks are inherently recurrent,causing information from distant steps to attenuate,they fail to effectively capture memory inconsistencies in practical offline learning schemes.Nonetheless,transplanting pretrained memory-consistent gate parameters into the LSTM model partially mitigates the performance deterioration caused by memory inconsistencies,suggesting that memory augmentation strategies have the potential to overcome this problem.As such a memory augmentation method,we propose the Gate-of-Gates(GoG)model,which extends the capacity of LSTM gates and demonstrates that it can mitigate additional errors arising from memory inconsistencies.展开更多
With the increasing growth of online news,fake electronic news detection has become one of the most important paradigms of modern research.Traditional electronic news detection techniques are generally based on contex...With the increasing growth of online news,fake electronic news detection has become one of the most important paradigms of modern research.Traditional electronic news detection techniques are generally based on contextual understanding,sequential dependencies,and/or data imbalance.This makes distinction between genuine and fabricated news a challenging task.To address this problem,we propose a novel hybrid architecture,T5-SA-LSTM,which synergistically integrates the T5 Transformer for semantically rich contextual embedding with the Self-Attentionenhanced(SA)Long Short-Term Memory(LSTM).The LSTM is trained using the Adam optimizer,which provides faster and more stable convergence compared to the Stochastic Gradient Descend(SGD)and Root Mean Square Propagation(RMSProp).The WELFake and FakeNewsPrediction datasets are used,which consist of labeled news articles having fake and real news samples.Tokenization and Synthetic Minority Over-sampling Technique(SMOTE)methods are used for data preprocessing to ensure linguistic normalization and class imbalance.The incorporation of the Self-Attention(SA)mechanism enables the model to highlight critical words and phrases,thereby enhancing predictive accuracy.The proposed model is evaluated using accuracy,precision,recall(sensitivity),and F1-score as performance metrics.The model achieved 99%accuracy on the WELFake dataset and 96.5%accuracy on the FakeNewsPrediction dataset.It outperformed the competitive schemes such as T5-SA-LSTM(RMSProp),T5-SA-LSTM(SGD)and some other models.展开更多
Lithium-ion batteries are the most widely accepted type of battery in the electric vehicle industry because of some of their positive inherent characteristics. However, the safety problems associated with inaccurate e...Lithium-ion batteries are the most widely accepted type of battery in the electric vehicle industry because of some of their positive inherent characteristics. However, the safety problems associated with inaccurate estimation and prediction of the state of health of these batteries have attracted wide attention due to the adverse negative effect on vehicle safety. In this paper, both machine and deep learning models were used to estimate the state of health of lithium-ion batteries. The paper introduces the definition of battery health status and its importance in the electric vehicle industry. Based on the data preprocessing and visualization analysis, three features related to actual battery capacity degradation are extracted from the data. Two learning models, SVR and LSTM were employed for the state of health estimation and their respective results are compared in this paper. The mean square error and coefficient of determination were the two metrics for the performance evaluation of the models. The experimental results indicate that both models have high estimation results. However, the metrics indicated that the SVR was the overall best model.展开更多
基金This work is supported by the National Key Research and Development Program of China(No.2023YFB4203000)the National Natural Science Foundation of China(No.U22A20178)
文摘Complicated loads encountered by floating offshore wind turbines(FOWTs)in real sea conditions are crucial for future optimization of design,but obtaining data on them directly poses a challenge.To address this issue,we applied machine learning techniques to obtain hydrodynamic and aerodynamic loads of FOWTs by measuring platform motion responses and wave-elevation sequences.First,a computational fluid dynamics(CFD)simulation model of the floating platform was established based on the dynamic fluid body interaction technique and overset grid technology.Then,a long short-term memory(LSTM)neural network model was constructed and trained to learn the nonlinear relationship between the waves,platform-motion inputs,and hydrodynamic-load outputs.The optimal model was determined after analyzing the sensitivity of parameters such as sample characteristics,network layers,and neuron numbers.Subsequently,the effectiveness of the hydrodynamic load model was validated under different simulation conditions,and the aerodynamic load calculation was completed based on the D'Alembert principle.Finally,we built a hybrid-scale FOWT model,based on the software in the loop strategy,in which the wind turbine was replaced by an actuation system.Model tests were carried out in a wave basin and the results demonstrated that the root mean square errors of the hydrodynamic and aerodynamic load measurements were 4.20%and 10.68%,respectively.
基金supported by the National Key Research and Development Program of China(No.2021YFB2600300).
文摘Intelligent maintenance of roads and highways requires accurate deterioration evaluation and performance prediction of asphalt pavement.To this end,we develop a time series long short-term memory(LSTM)model to predict key performance indicators(PIs)of pavement,namely the international roughness index(IRI)and rutting depth(RD).Subsequently,we propose a comprehensive performance indicator for the pavement quality index(PQI),which leverages the highway performance assessment standard method,entropy weight method,and fuzzy comprehensive evaluation method.This indicator can evaluate the overall performance condition of the pavement.The data used for the model development and analysis are extracted from tests on two full-scale accelerated test tracks,called MnRoad and RIOHTrack.Six variables are used as predictors,including temperature,precipitation,total traffic volume,asphalt surface layer thickness,pavement age,and maintenance condition.Furthermore,wavelet denoising is performed to analyze the impact of missing or abnormal data on the LSTM model accuracy.In comparison to a traditional autoregressive integrated moving average(ARIMAX)model,the proposed LSTM model performs better in terms of PI prediction and resiliency to noise.Finally,the overall prediction accuracy of our proposed performance indicator PQI is 93.8%.
文摘Deep learning plays a vital role in real-life applications, for example object identification, human face recognition, speech recognition, biometrics identification, and short and long-term forecasting of data. The main objective of our work is to predict the market performance of the Dhaka Stock Exchange (DSE) on day closing price using different Deep Learning techniques. In this study, we have used the LSTM (Long Short-Term Memory) network to forecast the data of DSE for the convenience of shareholders. We have enforced LSTM networks to train data as well as forecast the future time series that has differentiated with test data. We have computed the Root Mean Square Error (RMSE) value to scrutinize the error between the forecasted value and test data that diminished the error by updating the LSTM networks. As a consequence of the renovation of the network, the LSTM network provides tremendous performance which outperformed the existing works to predict stock market prices.
基金supported by the Open Fund of Magnetic Confinement Fusion Laboratory of Anhui Province(No.2024AMF04003)the Natural Science Foundation of Anhui Province(No.228085ME142)Comprehensive Research Facility for Fusion Technology(No.20180000527301001228)。
文摘This research focuses on solving the fault detection and health monitoring of high-power thyristor converter.In terms of the critical role of thyristor converter in nuclear fusion system,a method based on long short-term memory(LSTM)neural network model is proposed to monitor the operational state of the converter and accurately detect faults as they occur.By sampling and processing a large number of thyristor converter operation data,the LSTM model is trained to identify and detect abnormal state,and the power supply health status is monitored.Compared with traditional methods,LSTM model shows higher accuracy and abnormal state detection ability.The experimental results show that this method can effectively improve the reliability and safety of the thyristor converter,and provide a strong guarantee for the stable operation of the nuclear fusion reactor.
基金National Key R&D Program of China(No.2020YFB1707700)。
文摘The problems in equipment fault detection include data dimension explosion,computational complexity,low detection accuracy,etc.To solve these problems,a device anomaly detection algorithm based on enhanced long short-term memory(LSTM)is proposed.The algorithm first reduces the dimensionality of the device sensor data by principal component analysis(PCA),extracts the strongly correlated variable data among the multidimensional sensor data with the lowest possible information loss,and then uses the enhanced stacked LSTM to predict the extracted temporal data,thus improving the accuracy of anomaly detection.To improve the efficiency of the anomaly detection,a genetic algorithm(GA)is used to adjust the magnitude of the enhancements made by the LSTM model.The validation of the actual data from the pumps shows that the algorithm has significantly improved the recall rate and the detection speed of device anomaly detection,with the recall rate of 97.07%,which indicates that the algorithm is effective and efficient for device anomaly detection in the actual production environment.
基金supported by the National Natural Science Foundation of China (62003354)。
文摘This paper introduces the time-frequency analyzed long short-term memory(TF-LSTM) neural network method for jamming signal recognition over the Global Navigation Satellite System(GNSS) receiver. The method introduces the long shortterm memory(LSTM) neural network into the recognition algorithm and combines the time-frequency(TF) analysis for signal preprocessing. Five kinds of navigation jamming signals including white Gaussian noise(WGN), pulse jamming, sweep jamming, audio jamming, and spread spectrum jamming are used as input for training and recognition. Since the signal parameters and quantity are unknown in the actual scenario, this work builds a data set containing multiple kinds and parameters jamming to train the TF-LSTM. The performance of this method is evaluated by simulations and experiments. The method has higher recognition accuracy and better robustness than the existing methods, such as LSTM and the convolutional neural network(CNN).
文摘Load forecasting is of great significance to the development of new power systems.With the advancement of smart grids,the integration and distribution of distributed renewable energy sources and power electronics devices have made power load data increasingly complex and volatile.This places higher demands on the prediction and analysis of power loads.In order to improve the prediction accuracy of short-term power load,a CNN-BiLSTMTPA short-term power prediction model based on the Improved Whale Optimization Algorithm(IWOA)with mixed strategies was proposed.Firstly,the model combined the Convolutional Neural Network(CNN)with the Bidirectional Long Short-Term Memory Network(BiLSTM)to fully extract the spatio-temporal characteristics of the load data itself.Then,the Temporal Pattern Attention(TPA)mechanism was introduced into the CNN-BiLSTM model to automatically assign corresponding weights to the hidden states of the BiLSTM.This allowed the model to differentiate the importance of load sequences at different time intervals.At the same time,in order to solve the problem of the difficulties of selecting the parameters of the temporal model,and the poor global search ability of the whale algorithm,which is easy to fall into the local optimization,the whale algorithm(IWOA)was optimized by using the hybrid strategy of Tent chaos mapping and Levy flight strategy,so as to better search the parameters of the model.In this experiment,the real load data of a region in Zhejiang was taken as an example to analyze,and the prediction accuracy(R2)of the proposed method reached 98.83%.Compared with the prediction models such as BP,WOA-CNN-BiLSTM,SSA-CNN-BiLSTM,CNN-BiGRU-Attention,etc.,the experimental results showed that the model proposed in this study has a higher prediction accuracy.
基金The National Natural Science Foundation of China(No.61571106,61633013,61673108,81871444).
文摘To fully make use of information from different representation subspaces,a multi-head attention-based long short-term memory(LSTM)model is proposed in this study for speech emotion recognition(SER).The proposed model uses frame-level features and takes the temporal information of emotion speech as the input of the LSTM layer.Here,a multi-head time-dimension attention(MHTA)layer was employed to linearly project the output of the LSTM layer into different subspaces for the reduced-dimension context vectors.To provide relative vital information from other dimensions,the output of MHTA,the output of feature-dimension attention,and the last time-step output of LSTM were utilized to form multiple context vectors as the input of the fully connected layer.To improve the performance of multiple vectors,feature-dimension attention was employed for the all-time output of the first LSTM layer.The proposed model was evaluated on the eNTERFACE and GEMEP corpora,respectively.The results indicate that the proposed model outperforms LSTM by 14.6%and 10.5%for eNTERFACE and GEMEP,respectively,proving the effectiveness of the proposed model in SER tasks.
基金the Special Program for Innovation Methodology of the Ministry of Science and Technology of China(No.2016IM010100)。
文摘Load forecasting can increase the efficiency of modern energy systems with built-in measuring systerms by providing a more accurate peak power shaving performance and thus more reliable control.An analysis of an integrated CO2 heat pump and chiller system with a hot water storage system is presented in this paper.Drastic power fluctuations,which can be reduced with load forecasting,are found in historical operation records.A model that aims to forecast the ventilation system heating demand is thus established on the basis of a long short-term memory(LSTM)network.The model can successfully forecast the one hour ahead power using records of the past 48h of the system operation data and the ambient temperature.The mean absolute percentage error(MAPE)of the forecast results of the LSTM-based model is 10.70%,which is respectively 2.2%and 7.25%better than the MAPEs of the forecast results of the support vector regression based and persistence method based models.
基金the Open Fund of Guangxi Key Laboratory of Building New Energy and Energy Saving(Project Number:Guike Energy 17-J-21-3).
文摘Power prediction has been critical in large-scale wind power grid connections.However,traditional wind power prediction methods have long suffered from problems,for instance low prediction accuracy and poor reliability.For this purpose,a hybrid prediction model(VMD-LSTM-Attention)has been proposed,which integrates the variational modal decomposition(VMD),the long short-term memory(LSTM),and the attention mechanism(Attention),and has been optimized by improved dung beetle optimization algorithm(IDBO).Firstly,the algorithm's performance has been significantly enhanced through the implementation of three key strategies,namely the elite group strategy of the Logistic-Tent map,the nonlinear adjustment factor,and the adaptive T-distribution disturbance mechanism.Subsequently,IDBO has been applied to optimize the important parameters of VMD(decomposition layers and penalty factors)to ensure the best decomposition signal is obtained;Furthermore,the IDBO has been deployed to optimize the three key hyper-parameters of the LSTM,thereby improving its learning capability.Finally,an Attention mechanism has been incorporated to adaptively weight temporal features,thus increasing the model's ability to focus on key information.Comprehensive simulation experiments have demonstrated that the proposed model achieves higher prediction accuracy compared with VMD-LSTM,VMD-LSTM-Attention,and traditional prediction methods,and quantitative indexes verify the efectiveness of the algorithmic improvement as well as the excellence and precision of the model in wind power prediction.
基金supported by the National Natural Science Foundation of China(72061006)the research on the auxiliary diagnosis system of chronic injury of levator scapulae based on the concept of digital twin(Contract No:Qian Kehe Support[2023]General 117)Research on indoor intelligent assisted walking robot for the rehabilitation of walking ability of the elderly(Contract No:Qian kehe Support[2023]General 124).
文摘Stroke survivors often face significant challenges when performing daily self-care activities due to upper limb motor impairments.Traditional surface electromyography(sEMG)analysis typically focuses on isolated hand postures,overlooking the complexity of object-interactive behaviors that are crucial for promoting patient independence.This study introduces a novel framework that combines high-density sEMG(HD-sEMG)signals with an improved Whale Optimization Algorithm(IWOA)-optimized Long Short-Term Memory(LSTM)network to address this limitation.The key contributions of this work include:(1)the creation of a specialized HD-sEMG dataset that captures nine continuous self-care behaviors,along with time and posture markers,to better reflect real-world patient interactions;(2)the development of a multi-channel feature fusion module based on Pascal’s theorem,which enables efficient signal segmentation and spatial–temporal feature extraction;and(3)the enhancement of the IWOA algorithm,which integrates optimal point set initialization,a diversity-driven pooling mechanism,and cosine-based differential evolution to optimize LSTM hyperparameters,thereby improving convergence and global search capabilities.Experimental results demonstrate superior performance,achieving 99.58%accuracy in self-care behavior recognition and 86.19%accuracy for 17 continuous gestures on the Ninapro db2 benchmark.The framework operates with low latency,meeting the real-time requirements for assistive devices.By enabling precise,context-aware recognition of daily activities,this work advances personalized rehabilitation technologies,empowering stroke patients to regain autonomy in self-care tasks.The proposed methodology offers a robust,scalable solution for clinical applications,bridging the gap between laboratory-based gesture recognition and practical,patient-centered care.
基金supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea(NRF-2023S1A5A8077102).
文摘Deep learning enables neural networks to improve prediction performance through data supplementation.In financial time series forecasting,however,such data-driven approaches can encounter limitations where additional data degrade performance,contrary to common expectations.While more data can still be beneficial,it may introduce systemic concept drift due to the complex nonstationarities of stock price index time series,thereby exacerbating overfitting.One such drift is memory inconsistency:locally measured long memories fluctuate over time,alternately approaching and deviating from the random walk condition.We address this problem by typifying memory inconsistencies into two simplified forms:long-term dependentto-independent(D2I)and long-term independent-to-dependent(I2D)inconsistencies.The first experiment,which uses U.S.stock price indices,suggests that additional training examples may lead to performance deterioration of long short-term memory(LSTM)networks,especially when memory inconsistencies are prominent.Since stock markets are influenced by numerous unknown dynamics,the second experiment,which uses simulated mean-reverting time series derived from the fractional Ornstein–Uhlenbeck(fOU)process,is conducted to focus solely on challenges arising from memory inconsistencies.The experimental results demonstrate that memory inconsistencies disrupt the performance of LSTM networks.Theoretically,additional errors from D2I and I2D inconsistencies increase as the time lag increases.Since LSTM networks are inherently recurrent,causing information from distant steps to attenuate,they fail to effectively capture memory inconsistencies in practical offline learning schemes.Nonetheless,transplanting pretrained memory-consistent gate parameters into the LSTM model partially mitigates the performance deterioration caused by memory inconsistencies,suggesting that memory augmentation strategies have the potential to overcome this problem.As such a memory augmentation method,we propose the Gate-of-Gates(GoG)model,which extends the capacity of LSTM gates and demonstrates that it can mitigate additional errors arising from memory inconsistencies.
基金supported by Princess Nourah bint Abdulrahman University Researchers Supporting Project number(PNURSP2025R195)Princess Nourah bint Abdulrahman University,Riyadh,Saudi Arabia.
文摘With the increasing growth of online news,fake electronic news detection has become one of the most important paradigms of modern research.Traditional electronic news detection techniques are generally based on contextual understanding,sequential dependencies,and/or data imbalance.This makes distinction between genuine and fabricated news a challenging task.To address this problem,we propose a novel hybrid architecture,T5-SA-LSTM,which synergistically integrates the T5 Transformer for semantically rich contextual embedding with the Self-Attentionenhanced(SA)Long Short-Term Memory(LSTM).The LSTM is trained using the Adam optimizer,which provides faster and more stable convergence compared to the Stochastic Gradient Descend(SGD)and Root Mean Square Propagation(RMSProp).The WELFake and FakeNewsPrediction datasets are used,which consist of labeled news articles having fake and real news samples.Tokenization and Synthetic Minority Over-sampling Technique(SMOTE)methods are used for data preprocessing to ensure linguistic normalization and class imbalance.The incorporation of the Self-Attention(SA)mechanism enables the model to highlight critical words and phrases,thereby enhancing predictive accuracy.The proposed model is evaluated using accuracy,precision,recall(sensitivity),and F1-score as performance metrics.The model achieved 99%accuracy on the WELFake dataset and 96.5%accuracy on the FakeNewsPrediction dataset.It outperformed the competitive schemes such as T5-SA-LSTM(RMSProp),T5-SA-LSTM(SGD)and some other models.
文摘Lithium-ion batteries are the most widely accepted type of battery in the electric vehicle industry because of some of their positive inherent characteristics. However, the safety problems associated with inaccurate estimation and prediction of the state of health of these batteries have attracted wide attention due to the adverse negative effect on vehicle safety. In this paper, both machine and deep learning models were used to estimate the state of health of lithium-ion batteries. The paper introduces the definition of battery health status and its importance in the electric vehicle industry. Based on the data preprocessing and visualization analysis, three features related to actual battery capacity degradation are extracted from the data. Two learning models, SVR and LSTM were employed for the state of health estimation and their respective results are compared in this paper. The mean square error and coefficient of determination were the two metrics for the performance evaluation of the models. The experimental results indicate that both models have high estimation results. However, the metrics indicated that the SVR was the overall best model.
