Recurrent Neural Networks: A Comprehensive Review of Architectures, Variants, and Applications

Information, Год журнала: 2024, Номер 15(9), С. 517 - 517

Опубликована: Авг. 25, 2024

Recurrent neural networks (RNNs) have significantly advanced the field of machine learning (ML) by enabling effective processing sequential data. This paper provides a comprehensive review RNNs and their applications, highlighting advancements in architectures, such as long short-term memory (LSTM) networks, gated recurrent units (GRUs), bidirectional LSTM (BiLSTM), echo state (ESNs), peephole LSTM, stacked LSTM. The study examines application to different domains, including natural language (NLP), speech recognition, time series forecasting, autonomous vehicles, anomaly detection. Additionally, discusses recent innovations, integration attention mechanisms development hybrid models that combine with convolutional (CNNs) transformer architectures. aims provide ML researchers practitioners overview current future directions RNN research.

Язык: Английский

---

Deep Learning for Credit Card Fraud Detection: A Review of Algorithms, Challenges, and Solutions

IEEE Access, Год журнала: 2024, Номер 12, С. 96893 - 96910

Опубликована: Янв. 1, 2024

Deep learning (DL), a branch of machine (ML), is the core technology in today's technological advancements and innovations. learning-based approaches are state-of-the-art methods used to analyse detect complex patterns large datasets, such as credit card transactions. However, most fraud models literature based on traditional ML algorithms, recently, there has been rise applications deep techniques. This study reviews recent DL-based presents concise description performance comparison widely DL techniques, including convolutional neural network (CNN), simple recurrent (RNN), long short-term memory (LSTM), gated unit (GRU). Additionally, an attempt made discuss suitable metrics, common challenges encountered when training using architectures potential solutions, which lacking previous studies would benefit researchers practitioners. Meanwhile, experimental results analysis real-world dataset indicate robustness detection.

Язык: Английский

---

Optimized Ensemble Learning Approach with Explainable AI for Improved Heart Disease Prediction

Information, Год журнала: 2024, Номер 15(7), С. 394 - 394

Опубликована: Июль 8, 2024

Recent advances in machine learning (ML) have shown great promise detecting heart disease. However, to ensure the clinical adoption of ML models, they must not only be generalizable and robust but also transparent explainable. Therefore, this research introduces an approach that integrates robustness ensemble algorithms with precision Bayesian optimization for hyperparameter tuning interpretability offered by Shapley additive explanations (SHAP). The classifiers considered include adaptive boosting (AdaBoost), random forest, extreme gradient (XGBoost). experimental results on Cleveland Framingham datasets demonstrate optimized XGBoost model achieved highest performance, specificity sensitivity values 0.971 0.989 dataset 0.921 0.975 dataset, respectively.

Язык: Английский

---

A Comprehensive Review of Deep Learning: Architectures, Recent Advances, and Applications

Information, Год журнала: 2024, Номер 15(12), С. 755 - 755

Опубликована: Ноя. 27, 2024

Deep learning (DL) has become a core component of modern artificial intelligence (AI), driving significant advancements across diverse fields by facilitating the analysis complex systems, from protein folding in biology to molecular discovery chemistry and particle interactions physics. However, field deep is constantly evolving, with recent innovations both architectures applications. Therefore, this paper provides comprehensive review DL advances, covering evolution applications foundational models like convolutional neural networks (CNNs) Recurrent Neural Networks (RNNs), as well such transformers, generative adversarial (GANs), capsule networks, graph (GNNs). Additionally, discusses novel training techniques, including self-supervised learning, federated reinforcement which further enhance capabilities models. By synthesizing developments identifying current challenges, insights into state art future directions research, offering valuable guidance for researchers industry experts.

Язык: Английский

---

Deep learning-based dual optimization framework for accurate thyroid disease diagnosis using CNN architectures

Mehran University Research Journal of Engineering and Technology, Год журнала: 2025, Номер 44(2), С. 1 - 12

Опубликована: Апрель 9, 2025

Thyroid diseases, including hypothyroidism, hyperthyroidism, thyroid nodules, thyroiditis, and cancer, are among the most prevalent endocrine disorders, posing significant health risks, which need to be diagnosed treated promptly. Traditional diagnostic approaches, reliant on manual interpretation of medical images, time-consuming prone errors. This study introduces a novel deep learning framework utilizing advanced Convolutional Neural Networks (CNNs), specifically modified ResNet InceptionV3 architectures, improve accuracy efficiency disease diagnosis. We present Dual-OptNet, new hybrid architecture that effectively merges skip connections with multi-scale feature extraction based for lung classification tasks. Dual-OptNet shows accurate generalizability results in classifying an average best 97% from dual-step optimized using Adam SGD. Future work will focus developing real-time tool demonstrate potential utility this model clinical context. also enhancing dataset cover wider range uncommon cases, incorporating explainable AI methods, so decisions more interpretable. Further research explore ultrasound analysis multi-modal data integration, such as combining images patient history, enhance accuracy. Deploying system environments key validating its impact scalability, ultimately contributing efficient healthcare solutions

Язык: Английский

---

Effective Credit Risk Prediction Using Ensemble Classifiers With Model Explanation

IEEE Access, Год журнала: 2024, Номер 12, С. 115015 - 115025

Опубликована: Янв. 1, 2024

Язык: Английский

---

Analysis of thyroid nodule ultrasound images by image feature extraction technique

Современные инновации системы и технологии - Modern Innovations Systems and Technologies, Год журнала: 2024, Номер 4(3), С. 0301 - 0325

Опубликована: Сен. 11, 2024

The most frequent left thyroid nodule is the presence of nodules that have never been seen before. With X-ray computed tomography (CT) being used more often in diagnosing disorders, however, image processing has not applied frequently to standard machine learning due high density and artefacts found CT images gland. last section suggests a Convolutional Neural Network (CNN)-based end-to-end approach for automatic detection classification different types nodules. recommended model includes an improved segmentation network effectively divides regions within which each may be detected technique optimizes these areas. For example, 98% accuracy was obtained accurately categorising illness cases by examining aberrant modules X-rays. According our study, CNN can detect degrees severity caused located various parts body, thereby providing means through this procedure done automatically without requiring human intervention all time. Overall, study demonstrates how deep models identify diagnose using imaging, could increase precision effectiveness disease.

Язык: Английский

---

fNIRS Classification of Adults with ADHD Enhanced by Feature Selection

IEEE Transactions on Neural Systems and Rehabilitation Engineering, Год журнала: 2024, Номер 33, С. 220 - 231

Опубликована: Дек. 24, 2024

Adult attention deficit hyperactivity disorder (ADHD), a prevalent psychiatric disorder, significantly impacts social, academic, and occupational functioning. However, it has been relatively less prioritized compared to childhood ADHD. This study employed functional near-infrared spectroscopy (fNIRS) during verbal fluency tasks in conjunction with machine learning (ML) techniques differentiate between healthy controls (N=75) ADHD individuals (N=120). Efficient feature selection high-dimensional fNIRS datasets is crucial for improving accuracy. To address this, we propose hybrid method that combines wrapper-based embedded approach, termed Bayesian-Tuned Ridge RFECV (BTR-RFECV). The proposed facilitated streamlined hyperparameter tuning data, thereby reducing the number of features while enhancing HbO from combined frontal temporal regions were key, models achieving precision (89.89%), recall (89.74%), F-1 score (89.66%), accuracy MCC (78.36%), GDR (88.45%). outcomes this highlight promising potential combining ML as diagnostic tools clinical settings, offering pathway reduce manual intervention.

Язык: Английский

---

Early detection of thyroid disease using feature selection and hybrid machine learning approach

Deleted Journal, Год журнала: 2024, Номер 3

Опубликована: Дек. 30, 2024

In today's environment, thyroid disorders are quite widespread and widely dispersed. They frequently result in serious physical mental suffering. It interferes with the gland's ability to operate, which causes secrete too much hormone. The organs ground up by hormones produced when body enters auto-safe mode this illness. Avoiding condition is crucial because it has irreversible effects on body. Since disorder extremely difficult cure once reaches its final stage, preventing from occurring needs some awareness of development. ontological challenges disparate data standards that employed Medical Data Analysis (MDA) system-assisted healthcare management well-known industry. Rapid technological breakthroughs have drawn researchers health sector create accurate, dependable, reasonably priced medical (DSS) decision support systems (MDSS). Therefore, there continuous research being done construct an efficient practically applicable MFFN+MLP-based DSS for (MD) processing knowledge discovery (KD). Using computerised intelligent offers a practical way help professionals diagnose patients quickly correctly. Before diagnosis system can be created implemented, number problems must addressed handled, including how make decisions faced ambiguity imprecision.

Язык: Английский

---