Bidirectional Long Short-term Memory Research Articles

A Stock Prediction Method Based on Heterogeneous Bidirectional LSTM

A Stock Prediction Method Based on Heterogeneous Bidirectional LSTM

Impact Of Sarcasm Detection on Sentiment Analysis Using Bi-LSTM and FastText

Sentiment analysis categorizes a collection of texts in a document as either positive or negative. However, sometimes it cannot give accurate results due to sarcastic sentences. Sarcasm involves the use of positive language to convey negative meanings, So sarcasm detection is needed for sentiment classification to provide better results. One method that can be used to perform Sentiment classification is Bidirectional Long Short-Term Memory (Bi-LSTM). However, text data cannot be processed by Bi-LSTM, so it requires word embedding to convert text data into vectors. In this study, the word embedding used is FastText because it can learn the form of words by considering subword information. The results showed that sentiment classification with sarcasm detection could improve evaluation results by 0.08 for precision, 0.07 for recall, 0.07 for F1-score, and 0.07 for accuracy. A paired sample t-test was conducted on precision, recall, F1-score, and accuracy to examine whether there is a difference between sentiment classification with and without sarcasm detection. The obtained p-values are 2.84.10-9, 4.63.10-7, and 2.40.10-8, 6.22.10-8, respectively. This indicates a difference between sentiment classification with and without sarcasm detection. Therefore, with a 95% confidence level, it can be concluded that sarcasm detection impacts sentiment classification.

Hybrid Intelligent Anomaly Detection System Using Attention based Deep Learning Approach for Cyber Attacks Prevention

Objectives: Network Intrusion Detection System (NIDS) plays an important role in finding and preventing cyber-attacks, which helps to improve the entire security posture of an organization’s network infrastructure. The development of Deep Learning (DL) techniques possess the ability of IDS to detect attacks without delay and protects from intrusions even in real-time environment. Methods: The present study proposes an improved IDS framework called Enhanced Gated Recurrent Unit Hyper-Model combined Attention Bidirectional Long-Short Term Memory (EGHAB) approach, to effectively address the detection of attacks with maximum accuracy and minimal error rate. The proposed model is enhanced by using methods like numericalization and normalisation for pre-processing the input data, Genetic Algorithm (GA) for extracting intricate features from the input data, and the EGHAB classifier, which works on the principles of both Gated Recurrent Unit (GRU) and Bidirectional Long-Short Term Memory (BiLSTM) models along with attention mechanism. Findings: The EGHAB model efficiently learns the abnormal behaviour of network and classified the attack and non-attack data using the publicly available NSL-KDD dataset and a generated real time data set, scapy. And it achieved 99.94% accuracy over NSL KDD dataset during multiclass classification and 93.3% on real time data set with reduced Error rate. Additionally, to examine the efficacy, the proposed approach is compared with other existing methods and proved its improved performance. Novelty: A combined ensemble classifier with GRU, BiLSTM and attention mechanism is designed to predict the attacks in earlier stage rather than due over. The model achieved better accuracy and reduced false assumption using anomaly prediction mechanism. Keywords: Cyber Attack, Anomaly Detection, Deep learning, Gated Recurrent Unit, Bi­ Directional Long­-Short Term Memory, Attention mechanism

Leveraging advanced AI algorithms with transformer-infused recurrent neural networks to optimize solar irradiance forecasting

Solar energy (SE) is vital for renewable energy generation, but its natural fluctuations present difficulties in maintaining grid stability and planning. Accurate forecasting of solar irradiance (SI) is essential to address these challenges. The current research presents an innovative forecasting approach named as Transformer-Infused Recurrent Neural Network (TIR) model. This model integrates a Bi-Directional Long Short-Term Memory (BiLSTM) network for encoding and a Gated Recurrent Unit (GRU) network for decoding, incorporating attention mechanisms and positional encoding. This model is proposed to enhance SI forecasting accuracy by effectively utilizing meteorological weather data, handling overfitting, and managing data outliers and data complexity. To evaluate the model’s performance, a comprehensive comparative analysis is conducted, involving five algorithms: Artificial Neural Network (ANN), BiLSTM, GRU, hybrid BiLSTM-GRU, and Transformer models. The findings indicate that employing the TIR model leads to superior accuracy in the analyzed area, achieving R2 value of 0.9983, RMSE of 0.0140, and MAE of 0.0092. This performance surpasses those of the alternative models studied. The integration of BiLSTM and GRU algorithms with the attention mechanism and positional encoding has been optimized to enhance the forecasting of SI. This approach mitigates computational dependencies and minimizes the error terms within the model.

Joint intent detection and slot filling with syntactic and semantic features using multichannel CNN-BiLSTM

Understanding spoken language is crucial for conversational agents, with intent detection and slot filling being the primary tasks in natural language understanding (NLU). Enhancing the NLU tasks can lead to an accurate and efficient virtual assistant thereby reducing the need for human intervention and expanding their applicability in other domains. Traditionally, these tasks have been addressed individually, but recent studies have highlighted their interconnection, suggesting better results when solved together. Recent advances in natural language processing have shown that pretrained word embeddings can enhance text representation and improve the generalization capabilities of models. However, the challenge of poor generalization in joint learning models for intent detection and slot filling remains due to limited annotated datasets. Additionally, traditional models face difficulties in capturing both the semantic and syntactic nuances of language, which are vital for accurate intent detection and slot filling. This study proposes a hybridized text representation method using a multichannel convolutional neural network with three embedding channels: non-contextual embeddings for semantic information, part-of-speech (POS) tag embeddings for syntactic features, and contextual embeddings for deeper contextual understanding. Specifically, we utilized word2vec for non-contextual embeddings, one-hot vectors for POS tags, and bidirectional encoder representations from transformers (BERT) for contextual embeddings. These embeddings are processed through a convolutional layer and a shared bidirectional long short-term memory (BiLSTM) network, followed by two softmax functions for intent detection and slot filling. Experiments on the air travel information system (ATIS) and SNIPS datasets demonstrated that our model significantly outperformed the baseline models, achieving an intent accuracy of 97.90% and slot filling F1-score of 98.86% on the ATIS dataset, and an intent accuracy of 98.88% and slot filling F1-score of 97.07% on the SNIPS dataset. These results highlight the effectiveness of our proposed approach in advancing dialogue systems, and paving the way for more accurate and efficient natural language understanding in real-world applications.

Empowering Digital Civility with an NLP Approach for Detecting Twitter Cyberbullying through Boosted Ensembles

As the number of social networking sites grows, so do cyber dangers. Cyberbullying is harmful behaviour that uses technology to intimidate, harass or harm someone, often on social media platforms like Twitter. Machine learning is the optimal approach for cyberbullying detection on Twitter to process large amounts of data, identify patterns of offensive behaviour, and automate the detection process for corpus of tweets. To identify cyber threats using a trained model, the Boosted Ensemble (BE) technique is assessed with various machine learning algorithms such as Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), Naive Bayes (NB), Decision Tree (DT), Support Vector Machine (SVM), Bidirectional LSTM (BILSTM), Recurrent Neural Network LSTM (RNN-LSTM), Multi-Modal Cyberbullying Detection (MMCD) and Random Forest (RF). These classifiers are trained on the vectorized data to classify the tweets to identify cyberbullying threats. The proposed framework can detect cyberbullying cases precisely on tweets. The significance of the work lies in detecting and mitigating cyber threats in real-time and it impacts in enhancing the safety and well-being of social media users by reducing instances of cyberbullying and other cyber threat. The comparative analysis is done using metrices like accuracy, precision, recall and F1-score and the comparison results shows that boosted ensemble technique outperforms other compared algorithms with its overall performance. The accuracy rate of CNN, LSTM, NB, DT, SVM, RF, BILSTM and BE (92.5%, 93.5%, 84.6%, 88%, 89.3%, 92%, 93.75% and 96%), precision rate of CNN, LSTM, NB, DT, SVM, RF, RNN-LSTM and BE (90.2%, 91.3%, 88%, 85%, 86%, 91.6%, 92.1% and 94%), recall rate of CNN, LSTM, NB, DT, SVM, RF, BILSTM and BE (89.8%, 90.7%, 90%, 82%, 88.67%, 89%, 91.04% and 93.7%) and F1 score of CNN, LSTM, NB, DT, SVM, RF, MMCD and BE (90.6%, 91.8%, 85%, 84.56% 87.2%, 90%, 84.6% and 94.89%).

Prediction of Ductile Damage in Composite Material Used in Type IV Hydrogen Tanks by Artificial Neural Network and Machine Learning with Finite Element Modeling Approach

This study investigates the degradation process of composite materials used in high‐pressure hydrogen storage vessels by employing advanced computational techniques. A recurrent neural network, specifically a bidirectional long short‐term memory (Bi‐LSTM) network, is utilized to predict the temporal evolution of ductile damage. The key degradation features are extracted from finite element modeling (FEM) computations using group method of data handling algorithms and treated as time‐series data. Results demonstrate that the Bi‐LSTM network can accurately undergo both elastic and plastic behaviors of the composite under tensile strength. Additionally, traditional machine learning (ML) algorithms such as extreme gradient boosting and random forest are employed to forecast strain degradation, showing promising results. This hybrid approach combining FEM, ML, and deep learning provides a comprehensive method for predicting the degradation of composite materials, offering significant potential for optimizing the design and durability of hydrogen storage vessels.

Enhancing Breast Cancer Collagen Segmentation with Improved Swim Transformation Algorithm

The precise segmentation of collagen fibres in breast tissue is essential for an efficient diagnostic and treatment strategy because breast cancer is still a major worldwide health concern. By better separating fibers from the background, the Swim Transformation (SWT) algorithm has demonstrated potential to improve collagen fiber segmentation. Unfortunately, the resilience and accuracy of the original SWT method are limited because to its susceptibility to noise and distortions. This work suggests an updated version of the SWT algorithm that includes multiple improvements to maximize collagen segmentation in breast cancer tissues, thereby addressing these constraints. The novel method improves the detection of collagen fibers by introducing a noise reduction pre-processing step and adaptive thresholding. It also incorporates an extra post-processing phase to eliminate segmentation mistakes and improve the overall outcomes. Using a dataset of microscopic pictures of breast cancer tissue, the suggested method's effectiveness was assessed and contrasted with that of the original SWT algorithm and other cutting-edge segmentation techniques. The enhanced SWT algorithm achieves greater segmentation accuracy, sensitivity, and precision, according to experimental data. In addition, the study expands on previous research by using a deep learning framework that classifies breast cancer cells based on segmented collagen properties. This framework uses Long Short-Term Memory (LSTM) and Bidirectional LSTM models. The promise of deep learning for automated cancer detection was demonstrated when the model was trained on segmented collagen pictures to identify the presence of malignant cells. The suggested method provides a notable enhancement in the segmentation of collagen, yielding more dependable outcomes that aid medical professionals in diagnosing breast cancer and developing treatment plans.

Machine Learning Approach for Arabic Handwritten Recognition

Text recognition is an important area of the pattern recognition field. Natural language processing (NLP) and pattern recognition have been utilized efficiently in script recognition. Much research has been conducted on handwritten script recognition. However, the research on the Arabic language for handwritten text recognition received little attention compared with other languages. Therefore, it is crucial to develop a new model that can recognize Arabic handwritten text. Most of the existing models used to acknowledge Arabic text are based on traditional machine learning techniques. Therefore, we implemented a new model using deep machine learning techniques by integrating two deep neural networks. In the new model, the architecture of the Residual Network (ResNet) model is used to extract features from raw images. Then, the Bidirectional Long Short-Term Memory (BiLSTM) and connectionist temporal classification (CTC) are used for sequence modeling. Our system improved the recognition rate of Arabic handwritten text compared to other models of a similar type with a character error rate of 13.2% and word error rate of 27.31%. In conclusion, the domain of Arabic handwritten recognition is advancing swiftly with the use of sophisticated deep learning methods.

Optimized CNN-BiLSTM with Attention: A High Performance Model for Predicting Heart Disease Using Cleveland and Framingham Datasets

Worldwide, some 17.9 million survives are lost each year due to heart disease (HD), which is acknowledged by the World Health Organisation (WHO) as top cause of mortality. In order to simplify further action, HD prediction—a difficult problem—can give a computerised estimate of the HD level. Improving patient outcomes and allowing for timely medical interventions are both made possible by early detection and accurate calculation of HD. As a result, HD prediction has garnered a great deal of interest from healthcare facilities around the globe. There has been encouraging progress in the detection of cardiac illness thanks to recent developments in machine learning (ML). Transparency and explainability, in addition to generalisability and robustness, are crucial for ML models to be used in therapeutic settings. The efficient prediction and diagnosis of numerous diseases was greatly aided by systems based on Deep Learning (DL). By combining Convolutional Neural Networks (CNNs), Bidirectional Long Short-Term Memory (BiLSTMs), besides Attention Mechanisms (CNN-AM), this paper aims to build a strong HD prediction scheme. Minimal preparation is necessary for this procedure. To extract spatial features, CNN is used. To extract temporal characteristics, Bi-LSTM is used. Lastly, to filter out the outcomes of the more to ighted channel output classification, two channel to ights are allotted through the attention mechanism. The proposed model's parameters are fine-tuned using a new optimisation approach known as Newton-Raphson-based Optimiser (NRO), which ultimately leads to better classification accuracy. With accuracy of 95.3% on the Cleveland dataset and 98.1% on the Framingham dataset, respectively, the optimised CNN-BiLSTM-AM model demonstrated the best performance in the experimental findings.

Older Adult Fall Risk Prediction with Deep Learning and Timed Up and Go (TUG) Test Data.

Falls are a major health hazard for older adults; therefore, in the context of an aging population, predicting the risk of a patient suffering falls in the near future is of great impact for health care systems. Currently, the standard prospective fall risk assessment instrument relies on a set of clinical and functional mobility assessment tools, one of them being the Timed Up and Go (TUG) test. Recently, wearable inertial measurement units (IMUs) have been proposed to capture motion data that would allow for the building of estimates of fall risk. The hypothesis of this study is that the data gathered from IMU readings while the patient is performing the TUG test can be used to build a predictive model that would provide an estimate of the probability of suffering a fall in the near future, i.e., assessing prospective fall risk. This study applies deep learning convolutional neural networks (CNN) and recurrent neural networks (RNN) to build such predictive models based on features extracted from IMU data acquired during TUG test realizations. Data were obtained from a cohort of 106 older adults wearing wireless IMU sensors with sampling frequencies of 100 Hz while performing the TUG test. The dependent variable is a binary variable that is true if the patient suffered a fall in the six-month follow-up period. This variable was used as the output variable for the supervised training and validations of the deep learning architectures and competing machine learning approaches. A hold-out validation process using 75 subjects for training and 31 subjects for testing was repeated one hundred times to obtain robust estimations of model performances At each repetition, 5-fold cross-validation was carried out to select the best model over the training subset. Best results were achieved by a bidirectional long short-term memory (BLSTM), obtaining an accuracy of 0.83 and AUC of 0.73 with good sensitivity and specificity values.

Precision Offloading in Edge Computing: Leveraging Predictive Model

The intended effect of the investigation is to provide sophisticated prediction and decision-making models in order to optimize service delivery and improve the Quality of Experience (QoE) for users. This research tackles the problems that are associated with job offloading in edge computing settings. In order to reduce service latency and improve overall performance, the Bi-Directional Long Short-Term Memory (B-LSTM) model is used. This model provides the ability to forecast task creation and server load. In order to accommodate the particular qualities of different devices, the Selective Objective Offloading Decision (SOOD) approach is presented. This method makes use of the TOPSIS methodology to turn server assessment into a decision-making issue that involves several criteria. A considerable increase of 98.4% in user quality of experience is achieved by the SOOD paradigm. In addition, the Rapid Offloading Decision (ROD) model is presented in order to manage unexpected work patterns. This is accomplished by using the log information of surrounding devices, which results in instantaneous and dependable offloading choices. Through the usage of prediction algorithms and selective decision-making, this research gives a complete strategy to improving the efficiency of edge computing. The goal of this technique is to maximize the utilization of servers and the user experience.

Enhancing Photovoltaic Power Predictions with Deep Physical Chain Model

Predicting solar power generation is a complex challenge with multiple issues, such as data quality and choice of methods, which are crucial to effectively integrate solar power into power grids and manage photovoltaic plants. This study creates a hybrid methodology to improve the accuracy of short-term power prediction forecasts using a model called Transformer Bi-LSTM (Bidirectional Long Short-Term Memory). This model, which combines elements from the transformer architecture and bidirectional LSTM (Long–Short-Term Memory), is evaluated using two strategies: the first strategy makes a direct prediction using meteorological data, while the second employs a chain of deep learning models based on transfer learning, thus simulating the traditional physical chain model. The proposed approach improves performance and allows you to incorporate physical models to refine forecasts. The results outperform existing methods on metrics such as mean absolute error, specifically by around 24%, which could positively impact power grid operation and solar adoption.

An innovative attention infused- BiRecurrenTwin network assisted hybrid segmentation technique for accurate heart disease prediction

Heart disease is a noteworthy global health concern, profoundly affecting public health and individual well-being. Early prediction and diagnosis are crucial for reducing its negative impact. This work focuses on developing an effectual heart disease detection employing advanced optimization approaches and data processing. The proposed system incorporates a hybrid segmentation algorithm, Hybrid K-means-Fuzzy-C-means (HKMFCM) clustering, which assigns membership degrees to data points, enabling soft clustering and accommodating data uncertainty. Additionally, the Artificial Bee Colony (ABC) Optimization method is applied for feature selection, mimicking the foraging behavior of honeybees to identify the most discriminative attributes from the dataset. This optimization algorithm iteratively explores the feature space to select features that enhance the model's predictive accuracy. Furthermore, a novel classification architecture, termed the Attention-infused BiRecurrenTwin Network, is introduced to accurately predict heart disease based on segmented and extracted patient data profiles. This classifier leverages both Bidirectional Gated Recurrent Unit (BiGRU) and Bidirectional Long Short-Term Memory (BiLSTM) networks, with their bidirectional nature capturing both past and future contexts, thus increasing the classifier's capability to detect subtle temporal patterns in patient data. In addition, the proposed system addresses traditional approaches limitations through its advanced components like HKMFCM (effectively manages data uncertainty by enabling soft clustering), ABC Optimization technique (enabling an iterative, global search for the most discriminative attributes) Attention-infused BiRecurrenTwin Network (surpasses conventional classifiers by capturing both past and future temporal patterns in patient data). The simulation outcomes demonstrate that the developed system attains improved performance, with accuracy, ROC, F-measure, precision, recall, sensitivity, and specificity values of 96.09%, 97%, 95%, 96.6%, 94.3%, 95.14%, and 97.05%, respectively which is higher than the baseline models with an average of 7.75 % increase in accuracy. These results indicate that the developed predictive models show promise in accurately classifying individuals based on their extracted features, thereby facilitating the early recognition of heart disease.

Deep learning-powered visual place recognition for enhanced mobile multimedia communication in autonomous transport systems

The progress of autonomous transport systems (ATS) involves efficient multimedia communication for real-time data tradeoffs and environmental issues. Deep learning (DL) powered visual place recognition (VPR) was developed as an effective tool to improve mobile multimedia communication in ATS. VPR relates to the capability of a method or device to recognize and identify particular places or locations from the visual scene. This procedure involves inspecting visual data, like images or video frames, to control the unique features or features connected with diverse locations. By leveraging camera sensors, VPR allows vehicles to detect their surroundings, enabling context-aware communication and enhancing the entire system's performance. DL-empowered VPR offers a transformative manner to improve mobile multimedia communication in ATS. By identifying and understanding their situation, autonomous vehicles can communicate most effectively and operate reliably and safely, paving the way for a future characterized by seamless and intelligent transportation. This article develops a novel Deep Learning-Powered Visual Place Recognition for Enhanced Multimedia Communication in Autonomous Transport Systems (DLVPR-MCATS) methodology. The main aim of the DLVPR-MCATS methodology is to recognize visual places or not utilize optimal DL approaches. For this purpose, the DLVPR-MCATS approach utilizes a bilateral filtering (BF) based preprocessing model. For the feature fusion model, the DLVPR-MCATS approach follows three models: residual network (ResNet), EfficientNet, and MobileNetv2. Moreover, the hyperparameter tuning method uses the Harris Hawks Optimization (HHO) model. Finally, the bidirectional long short-term memory (BiLSTM) technique is implemented to recognize visual places. A wide range of simulations is executed to validate the solution of the DLVPR-MCATS method. The experimental validation of the DLVPR-MCATS method portrayed a superior performance over other models concerning various aspects.

MC-VMD-CNN-BiLSTM short-term wind power prediction considering rolling error correction

Abstract Wind energy is a clean and renewable source that has the potential to alleviate the global fossil fuel crisis and environmental pollution by generating electricity. However, accurately predicting wind energy output remains challenging due to its inherent uncertainty. To enhance the accuracy of wind power prediction, a short-term wind power forecasting method for power systems, MC-VMD-CNN-BiLSTM, is proposed, which considers error rolling correction. The method begins with feature selection and outlier handling using the quadrature method. Then, wind power data is decomposed into multiple sub-sequences using the Variational Mode Decomposition (VMD) technique to reduce the raw volatility of wind power. Then, a Convolutional Neural Network (CNN) followed by a Bidirectional Long Short-Term Memory (BiLSTM) model is used for wind power prediction. Finally, the proposed method utilizes the Monte Carlo method for rolling error correction by using known errors from previous time frames to correct subsequent predictions. The MC-VMD-CNN-BiLSTM proposed in this paper considering error rolling correction is compared with ELM, SVM, PSO-BP and ARIMA models through an example analysis of the data of a city, and the proposed model in this paper reduces 61.78%, 50.35%, 62.30% and 73.05% in the NRMSE index in the spring as an example, respectively. The results show that the prediction model proposed in this paper has higher prediction accuracy compared with the traditional prediction model.

Dual-model approach for one-shot lithium-ion battery state of health sequence prediction

Lithium-ion batteries play a crucial role in powering various applications, including Electric Vehicles (EVs), underscoring the importance of accurately estimating their State Of Health (SOH) throughout their operational lifespan. This paper introduces two novel models: a Transformer (TOPS-SoH) and a Long Short-Term Memory based (LSTM-OSoH) for One-shot Prediction of SOH. The LSTM-OSoHexcels in accuracy, achieving a Masked Mean Absolute Error (MMAE) of less than 0.01 for precise SOH estimation, while the TOPS-SoHdemonstrates simplicity and efficiency, with accuracy comparable to state-of-the-art models. The TOPS-SoHmodel also offers additional interpretability by providing insights into the attention scores between inputs and outputs, highlighting the cycles used for estimation. These models were trained using the MIT battery dataset, with auto-encoders employed to reduce the dimensionality of the input data. Additionally, the models’ effectiveness was validated against a Bidirectional LSTM (BiLSTM) baseline, demonstrating superior performance in terms of lower MMAE, MMSE, and MAPE values, making them highly suitable for integration into Battery Management Systems (BMS). These findings contribute to advancing SOH estimation up to the End Of Life (EOL), which is crucial for ensuring the reliability and longevity of lithium-ion batteries in diverse applications.

Enhancing solar irradiance forecasting for hydrogen production: The MEMD-ALO-BiLSTM hybrid machine learning model

This study focuses on an innovative hybrid machine-learning model for solar irradiance forecasting, targeting the integration of solar power into hydrogen production systems. Addressing the urgent need for sustainable energy transitions, the paper introduces the MEMD-ALO-BiLSTM model, designed to enhance solar irradiance prediction accuracy. This model uniquely combines Multivariate Empirical Mode Decomposition (MEMD), Ant Lion Optimizer (ALO), and Bidirectional Long Short-Term Memory (BiLSTM) techniques, setting a new benchmark in forecast precision across various seasonal datasets from Jiangsu Province, China. Demonstrating superior performance to traditional models, it achieves an exceptional coefficient of determination, averaging 0.99 for all seasons. Additionally, to prove the efficiency of the model three statistical tests were used, namely Wilcoxon, Friedman, and P-value. The research highlights the model's potential in optimizing photovoltaic systems and hydrogen production, thus contributing to carbon dioxide emission mitigation. Through comprehensive simulations of a residential system encompassing photovoltaic cells, compressors, and electrolyzers, the study underscores the practical feasibility and significant advancements the MEMD-ALO-BiLSTM model offers in the renewable energy sector, promoting a shift toward more reliable and efficient solar-powered hydrogen generation systems. Accordingly, the day-ahead values of photovoltaic-generated power and hydrogen production through the electrolyzer reached peak values at 1:00PM with approximately 75 kW and 1.4 kg, respectively.

Generative AI with WGAN-GP for boosting seizure detection accuracy.

Imbalanced datasets pose challenges for developing accurate seizure detection systems based on electroencephalogram (EEG) data. Generative AI techniques may help augment minority class data to facilitate automatic epileptic seizure detection. This study investigates the impact of various data augmentation (DA) approaches, including Wasserstein Generative Adversarial Network with Gradient Penalty (WGAN-GP), Vanilla GAN, Conditional GAN (CGAN), and Cramer GAN, on classification performance with Random Forest models. The best-performing GAN variant, WGAN-GP, was then integrated with a bidirectional Long Short-Term Memory (LSTM) architecture and compared against traditional and synthetic oversampling methods. The evaluation of different GAN variants for data augmentation with Random Forest classifiers identified WGAN-GP as the most effective approach. The integration of WGAN-GP with bidirectional LSTM yielded substantial performance improvements, outperforming traditional oversampling methods and achieving an accuracy of 91.73% on the augmented data, compared to 86% accuracy on real data without augmentation. The proposed generative AI approach combining WGAN-GP and recurrent neural network models outperforms comparative synthetic oversampling methods on metrics relevant for reliable seizure detection from imbalanced EEG datasets. Incorporating the WGAN-GP generative AI technique for data augmentation and integrating it with bidirectional LSTM elevates seizure detection accuracy for imbalanced EEG datasets, surpassing the performance of traditional oversampling and class weight adjustment methods. This approach shows promise for improving epilepsy monitoring and management through enhanced automated detection system effectiveness.

Vessel Trajectory Prediction at Inner Harbor Based on Deep Learning Using AIS Data

This study aims to improve vessel trajectory prediction in the inner harbor of Busan Port using Automatic Identification System (AIS) data and deep-learning techniques. The research addresses the challenge of irregular AIS data intervals through linear interpolation and focuses on enhancing the accuracy of predictions in complex port environments. Recurrent neural network (RNN), Long Short-Term Memory (LSTM), Bidirectional LSTM, Gated Recurrent Unit (GRU), and Bidirectional GRU models were developed, with LSTM delivering the highest performance. The primary scientific question of this study is how to reliably predict vessel trajectories under varying conditions in inner harbors. The results demonstrate that the proposed method not only improves the precision of predictions but also identifies critical areas where Vessel Traffic Service Operators (VTSOs) can better manage vessel movements. These findings contribute to safer and more efficient vessel traffic management in ports with high traffic density and complex navigational challenges.