A Deep Learning Architecture for Spatio-Temporal Feature Extraction and Alcoholism Detection

In this paper, we propose an emotion recognition model based on convolutional neural network (CNN), long short term memory (LSTM) and channel attention mechanism, aiming at the low classification accuracy of machine learning methods and the uneven spatial distribution of electroencephalogram (EEG) electrodes. This model can effectively integrate the frequency, space and time information of EEG signals, and improve the accuracy of emotion recognition by adding channel attention mechanism after the last convolutional layer of the model. Firstly, construct a 4-dimensional structure representing EEG signals. Then, a CLSTM model structure combining CNN and LSTM is designed. CNN is used to extract frequency and spatial information from 4-dimensional input, and LSTM is used to extract time information. Finally, the channel attention module is added after the last convolutional layer of CLSTM model structure to allocate the weight of different electrodes. In this paper, an emotion recognition model based on CLSTM and channel attention mechanism was proposed from the perspective of integrating the frequency, space and time 3-dimensional information of EEG signals. The average classification accuracy of the model on SEED public data set reached 93.36%, which was significantly improved over the existing CNN and LSTM emotion recognition models.

Automated accurate emotion recognition system using rhythm-specific deep convolutional neural network technique with multi-channel EEG signals

Emotion recognition (ER) is crucial for understanding human behaviours, social interactions, and psychological well-being. Electroencephalography (EEG) has emerged as a promising tool for capturing the neural correlates of emotions. This work is a systematic review of articles in ER using EEG signals. A total of 120 articles from 1041 articles were selected based on PRISMA guidelines using defined inclusion and exclusion criteria, published between 2018 and 2024. This article aims to provide an in-depth understanding of the current landscape of ER from EEG signals utilizing deep learning (DL). This review offers valuable guidance for researchers and practitioners seeking more refined and reliable emotion classification systems. To explore the effectiveness of DL models in EEG-based ER, several potential DL models, such as convolutional neural network, long short-term memory (LSTM), gated recurrent unit (GRU), hybrid bidirectional LSTM (BiLSTM), bidirectional GRU, and advanced DL models such as convolutional recurrent neural network and EEG-Conformer models are applied to two popular datasets, SEED and GAMEEMO, respectively, to depict the full process of ER. Additionally, the performance of DL models is also compared with the performance of basic machine learning (ML) models such as SVM, k-nearest neighbors, logistic regression, and boosting algorithms such as AdaBoost, XGBoost and LightGBM. Through extensive experiments and performance evaluations, the performance of different models when applied to the datasets mentioned above is compared. The accuracy, precision, recall, and F1-scores are analysed to determine the most effective model for EEG-based ER. The findings of this study demonstrate that the performance of hybrid DL models is more efficacious than that of ML models. The best-performing model (BiLSTM) classified the emotions, with an accuracy of 90.54% when applied to the GAMEEMO dataset. This research contributes to the growing body of literature on ER and provides insights into the feasibility of using EEG signals to understand emotional states, and presents a structured roadmap for future exploration. The findings can aid in the development of more accurate and reliable ER systems, which can have wide-ranging applications in psychology, social sciences, and human-computer interactions.

Wearable electroencephalography (EEG) devices for seizure detection accuracy and reliability are deep learning (DL) applications in the field of epilepsy diagnosis. In this study, we sought to increase the accuracy of seizure detection using advanced DL algorithms on the Children’s Hospital Boston - Massachusetts Institute of Technology (CHB-MIT) EEG database. First, a fully convolutional network (FCN) was trained and assessed using accuracy and recall/precision metrics, and the early stopping technique was used to avoid overfitting. To assess the performance, the FCN was evaluated in terms of various metrics, including accuracy, precision, recall, F1-score, and receiver operating characteristic (ROC)-area under the curve (AUC). In addition, two-dimensional (2D) convolutional neural networks (CNNs) and long short-term memory (LSTM) models were used to model the database, and their performance was thoroughly measured using different metrics, graphs, and confusion matrices. Using LSTM variants, such as standard LSTM, bidirectional LSTM, stacked LSTM, and LSTM attention mechanisms, hybrid convolutional LSTM (ConvLSTM) models were trained and compared. The comparison was conducted based on the training and validation accuracy and loss, as well as the graphs resulting from the precision–recall curves. Apart from DL approaches, EEG signal analysis using time–frequency techniques, such as wavelet transform and short-time Fourier transform, has also been investigated. These methods assisted in the analysis of the time–frequency features of EEG signals in combination with DL models. This study demonstrates that the performance of wearable EEG devices can be augmented using a combination of DL and seizure signal processing techniques. The FCN achieved an accuracy of 92%, with a recall for seizures of 33%, an F1-score of 0.03, and strong ROC-AUC results. The 2D CNN achieved 96% accuracy, a seizure recall of 70%, an F1-score of 0.12, and an ROC-AUC score of 78%. The baseline LSTM struggled with effectiveness at 53% accuracy with a seizure recall of 18%. In contrast, the LSTM model, which incorporated synthetic minority oversampling technique (SMOTE) balancing, was able to reach improvements of up to 89% accuracy, with a precision of 91%, a recall of 86%, an F1-score of 0.89, and a strong ROC curve performance. Among the models, the LSTM with SMOTE was the best performer, with 89% accuracy, 91% precision, 86% recall, and an F1-score of 0.89. These results provide evidence that applying techniques for data balancing in combination with certain DL network architectures significantly improves the detection of seizures using wearable EEG devices worn on the body. We believe that real-time monitoring and high-performance systems are feasible using optimized DL frameworks. The analysis of the performance of different models allows for the understanding of the possibilities of optimizing the architectures of DL algorithms for the modern diagnosis of epilepsy in real time. The source code used to carry out the experiments is publicly available at CHB-MIT EEG Dataset Python Scripts (https://www.kaggle.com/code/adnankust/adnaneeg1).

Online leakage current classification using convolutional neural network long short-term memory for high voltage insulators on web-based service

IntroductionDrug safety has increasingly become a serious public health problem that threatens health and damages social economy. The common detection methods have the problem of high false positive rate. This study aims to introduce deep learning models into the adverse drug reaction (ADR) signal detection and compare different methods.MethodsThe data are based on adverse events collected by Center for ADR Monitoring of Guangdong. Traditional statistical methods were used for data preliminary screening. We transformed data into free text, extracted text information and made classification prediction by using the Long Short-Term Memory (LSTM) model. We compared it with the existing signal detection methods, including Logistic Regression, Random Forest, K-NearestNeighbor, and Multilayer Perceptron. The feature importance of the included variables was analyzed.ResultsA total of 2,376 ADR signals were identified between January 2018 and December 2019, comprising 448 positive signals and 1,928 negative signals. The sensitivity of the LSTM model based on free text reached 95.16%, and the F1-score was 0.9706. The sensitivity of Logistic Regression model based on feature variables was 86.83%, and the F1-score was 0.9063. The classification results of our model demonstrate superior sensitivity and F1-score compared to traditional methods. Several important variables “Reasons for taking medication”, “Serious ADR scenario 4”, “Adverse reaction analysis 5”, and “Dosage” had an important influence on the result.ConclusionThe application of deep learning models shows potential to improve the detection performance in ADR monitoring.

The rapid evolution of healthcare technologies has spurred a shift towards continuous, real-time monitoring of patients' vital signs, creating an urgent need for advanced anomaly detection systems. This chapter explores the application of deep learning models, particularly Autoencoders and Long Short-Term Memory (LSTM) networks, in the detection of anomalies within vital sign data streams. By leveraging the unique strengths of these models, healthcare systems can effectively identify subtle deviations in patient health metrics, such as heart rate, blood pressure, and respiratory rate, that may indicate deteriorating conditions. Autoencoders, with their ability to learn normal patterns from unlabeled data, and LSTMs, designed to capture long-term temporal dependencies in time-series data, are presented as powerful tools for addressing the challenges associated with real-time anomaly detection in clinical settings. Through case studies and practical applications, this chapter highlights the significant potential of these deep learning techniques in improving patient outcomes through early intervention, enhanced prediction of acute events, and more personalized care strategies. The growing integration of these models in wearable devices, mobile health applications, and remote patient monitoring systems is also examined, emphasizing their role in facilitating proactive and patient-centered healthcare. Key challenges related to data noise, missing values, and scalability are addressed, alongside strategies for overcoming these limitations in real-world healthcare environments. This chapter contributes valuable insights into the transformative potential of machine learning in modern healthcare, providing a foundation for future research and development in anomaly detection, health monitoring, and predictive healthcare analytics.

Accurate pollutant prediction is essential in fields such as meteorology, meteorological disasters, and climate change studies. In this study, long short-term memory (LSTM) and deep neural network (DNN) models were applied to six pollutants and comprehensive air-quality index (CAI) predictions from 2015 to 2020 in Korea. In addition, we used the network method to find the best data sources that provide factors affecting comprehensive air-quality index behaviors. This study had two steps: (1) predicting the six pollutants, including fine dust (PM10), fine particulate matter (PM2.5), ozone (O3), sulfurous acid gas (SO2), nitrogen dioxide (NO2), and carbon monoxide (CO) using the LSTM model; (2) forecasting the CAI using the six predicted pollutants in the first step as predictors of DNNs. The predictive ability of each model for the six pollutants and CAI prediction was evaluated by comparing it with the observed air-quality data. This study showed that combining a DNN model with the network method provided a high predictive power, and this combination could be a remarkable strength in CAI prediction. As the need for disaster management increases, it is anticipated that the LSTM and DNN models with the network method have ample potential to track the dynamics of air pollution behaviors.

Detecting emotions helps create more intuitive and responsive technology interfaces between humans and machines. This study introduces deep learning-based emotion detection using EEG signals. Our research uses Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM) networks, and a proposed hybrid of CNN and LSTM (ConvLSTM) along with Particle Swarm Optimization (PSO) for hyperparameter optimization. The presented study aims to improve EEG-based emotion detection, which will help create more unified and emotionally intelligent human-machine interfaces. Recent deep learning advances have advanced emotion detection. These advances have allowed systems to understand and respond to human emotions. Electroencephalography (EEG) signals can capture neural patterns linked to different emotions, giving a unique view of the brain's emotional state. However, extracting meaningful emotional information from these complex signals is difficult. EEG spatial patterns are captured using CNN and LSTM networks model data temporal dependencies. A hybrid ConvLSTM model integrates spatial and temporal information for a complete EEG analysis. The study presents a method that helps to detect emotions from EEG signals with accuracy of 99.63%. This study's high accuracy is a significant advancement in the field and has great potential for improving human-machine interaction systems. Due to the use of deep learning-based emotion detection, these systems are able to intuitively respond to the feelings of their users. They allow for deeper and more genuine connections to be made between humans and machines. The presented methodology achieves remarkable precision and lays the groundwork for emotionally intelligent technologies that can impact virtual assistants, healthcare, and entertainment. This research advances emotionally intelligent human-machine interfaces, paving the way for technology to understand and respond to human emotions with precision and sensitivity.

Freezing of Gait (FOG) is an impairment that affects the majority of patients in the advanced stages of Parkinson’s Disease (PD). FOG can lead to sudden falls and injuries, negatively impacting the quality of life for the patients and their families. Rhythmic Auditory Stimulation (RAS) can be used to help patients recover from FOG and resume normal gait. RAS might be ineffective due to the latency between the start of a FOG event, its detection and initialization of RAS. We propose a system capable of both FOG prediction and detection using signals from tri-axial accelerometer sensors that will be useful in initializing RAS with minimal latency. We compared the performance of several time frequency analysis techniques, including moving windows extracted from the signals, handcrafted features, Recurrence Plots (RP), Short Time Fourier Transform (STFT), Discreet Wavelet Transform (DWT) and Pseudo Wigner Ville Distribution (PWVD) with Deep Learning (DL) based Long Short Term Memory (LSTM) and Convolutional Neural Networks (CNN). We also propose three Ensemble Network Architectures that combine all the time frequency representations and DL architectures. Experimental results show that our ensemble architectures significantly improve the performance compared with existing techniques. We also present the results of applying our method trained on a publicly available dataset to data collected from patients using wearable sensors in collaboration with A.T. Still University.

This study compared popular Deep Learning (DL) architectures to classify machining surface roughness using sound and force data. The DL architectures considered in this study include Multi-Layer Perceptron (MLP), Convolution Neural Network (CNN), Long Short-Term Memory (LSTM), and transformer. The classification was performed on the sound and force data generated during machining aluminum sheets for different levels of spindle speed, feed rate, depth of cut, and end-mill diameter, and it was trained on 30 s machining data (10–40 s) of the machining experiments. Since a raw audio waveform is seldom used in DL models, Mel-Spectrogram and Mel Frequency Cepstral Coefficients (MFCCs) audio feature extraction techniques were used in the DL models. The results of DL models were compared for the training–validation accuracy, training epochs, and training parameters of each model. Although the roughness classification by all the DL models was satisfactory (except for CNN with Mel-Spectrogram), the transformer-based modes had the highest training (>96%) and validation accuracies (≈90%). The CNN model with Mel-Spectrogram exhibited the worst training and inference accuracy, which is influenced by limited training data. Confusion matrices were plotted to observe the classification accuracy visually. The confusion matrices showed that the transformer model trained on Mel-Spectrogram and the transformer model trained on MFCCs correctly predicted 366 (or 91.5%) and 371 (or 92.7%) out of 400 test samples. This study also highlights the suitability and superiority of the transformer model for time series sound and force data and over other DL models.

Due to over exploitation of fossil fuels, many governments have started promoting renewable energy sources and particularly wind energy is emerging as the prominent replacement to the conventional electricity generators. The limited land availability and issues pertaining to it are making the way forward for offshore wind energy. Wind speed forecasting [14] helps in optimal planning [1] [10] of the generation and demand. It also makes evaluating monthly revenues easier and reduces unexpected shut down of wind turbines by Grid operators during windy seasons avoiding revenue loss to the investors. Forecasting wind speed using statistical model is common. With the recent rise in application of deep learning models [9], state of the art approaches are able to produce forecasting results with promising accuracy compared to the traditional time series models. This work focuses on short term wind speed forecasting at offshore wind site using deep learning approaches such as Long Short Term Memory (LSTM) [6] and Gated Recurrent Unit (GRU) [2] models.

In this study, we investigate various deep learning models based on convolutional neural networks (CNNs) and Long Short Term Memory (LSTM) recurrent neural networks for sentiment analysis of Arabic microblogs. Unlike English, the Arabic language has several specifics which complicate the process of feature extraction by traditional methods. We adopted a neural language model created at Google, known as word2vec, for vectorizing text. We then designed and evaluated several deep learning architectures using CNN and LSTM. The experiments were run on two publicly available Arabic tweets datasets. Promising results have been attained when combining LSTMs and compared favorably with most related work.

While electroencephalography (EEG)-based user authentication has demonstrated strong potential in controlled laboratory conditions, its reliability in more realistic settings remains underexplored. In this work, we investigate EEG-based user authentication in realistic interactive scenarios involving cognitively and behaviorally rich tasks. We collected EEG data from subjects under two conditions: solo reading aloud and two-person dialog-based games. These scenarios represent practical, everyday activities where users produce speech and engage in turn-taking interactions, introducing non-stationarity, muscle artifacts, and attention shifts that challenge conventional EEG-based models. To evaluate the user authentication performance under these challenges, we systematically apply various deep learning backbone models and loss functions under a fair and consistent experimental protocol. Specifically, we first divide the whole EEG signal into short windows with pre-defined length and stride. For each window, a time-frequency representation is computed using a continuous wavelet transform, resulting in a sequence of 2D time-frequency maps that capture the non-stationary characteristics of EEG signals. These maps are then fed into various backbone models, such as long short term memory (LSTM) and convolutional neural network (CNN), which are trained to extract robust identity-discriminative features. To further enhance inter-subject separability, we employ the Softmax loss or ArcFace loss during training. Experimental results demonstrate that even under dynamic and less controlled conditions, EEG signals retain individual-specific patterns, yielding high authentication accuracy. These findings highlight the feasibility of extending EEG-based biometrics to more natural environments.

The prediction of stock price movement direction is significant in financial circles and academic. Stock price contains complex, incomplete, and fuzzy information which makes it an extremely difficult task to predict its development trend. Predicting and analysing financial data is a nonlinear, time-dependent problem. With rapid development in machine learning and deep learning, this task can be performed more effectively by a purposely designed network. This paper aims to improve prediction accuracy and minimizing forecasting error loss through deep learning architecture by using Generative Adversarial Networks. It was proposed a generic model consisting of Phase-space Reconstruction (PSR) method for reconstructing price series and Generative Adversarial Network (GAN) which is a combination of two neural networks which are Long Short-Term Memory (LSTM) as Generative model and Convolutional Neural Network (CNN) as Discriminative model for adversarial training to forecast the stock market. LSTM will generate new instances based on historical basic indicators information and then CNN will estimate whether the data is predicted by LSTM or is real. It was found that the Generative Adversarial Network (GAN) has performed well on the enhanced root mean square error to LSTM, as it was 4.35% more accurate in predicting the direction and reduced processing time and RMSE by 78 secs and 0.029, respectively. This study provides a better result in the accuracy of the stock index. It seems that the proposed system concentrates on minimizing the root mean square error and processing time and improving the direction prediction accuracy, and provides a better result in the accuracy of the stock index.