Search
61 articles for “LSTM”
-
A Comparative Study of Deep Learning Methods for Depression Detection in Social Media Data
Abstract: With the rise of social media platforms like Twitter, Reddit, and Facebook, individuals increasingly share personal information about their moods, behaviors, and mental states. This trend provides a unique opportunity to leverage large-scale textual data for understanding and monitoring mental health conditions, particularly depression, a prevalent and challenging mental health issue. Traditional depression assessments are often confined to clinical environments and lack the capacity for real-time monitoring. In contrast, social …
Published in Journal of Artificial Intelligence Research & Advances · Vol. 12, Issue 3, 2025 · pp. 55–65 Read article
-
Forecasting Commodity Prices Using Deep Learning Techniques: An Empirical Evidence from India
Abstract: Commodity price forecasting is instrumental in financial markets, providing framework for investment choices and risk management practices. Traditional models, including statistical and machine learning approaches, have limitations in capturing the nonlinear and volatile nature of commodity prices. Deep learning (DL) techniques have emerged as promising alternatives, leveraging advanced neural networks to enhance predictive accuracy. This study presents a thorough and comprehensive examination of deep learning applications in commodity price prediction, …
Published in Journal of Computer Technology & Applications · Vol. 16, Issue 3, 2025 · pp. 08–12 Read article
-
Comparative Analysis of AI-Based Approach vs. Traditional Methods in Climate Modeling
Abstract: Climate modeling helps to predict the future of climate variations and human interference with environment. The traditional General Circulation Models (GCMs) are based on physics-derived mathematical equations but are very expensive in terms of computation. There are alternative ways to perform climate modeling in recent years with the rise and improvement of Artificial Intelligence (AI) based approaches in term of predictability, efficiency, and classification of extreme events compared to conventional. …
Published in Current Trends in Information Technology · Vol. 15, Issue 3, 2025 · pp. 26–32 Read article
-
AI-Driven Micro-Expression Recognition for Early Mental Health Disorder
Abstract: Mental health conditions like anxiety and depression are often undiagnosed because the usual diagnostic methods based on basic regular instruments like questionnaires and clinical interviews have some limitations in them. They are not objective often and may not catch the initial signs of psychological distress. Micro-expressions have become valid measures of repressed or unconscious emotions and can provide greater insight into someone's mental condition. Also, identification and interpretation of these …
Published in Journal of Mobile Computing, Communications & Mobile Networks · Vol. 12, Issue 3, 2025 · pp. 40–49 Read article
-
Deep Learning-Enhanced Polymer-Based Wearable Biosensors for Continuous Health Tracking via IoT
Abstract: The rapid proliferation of wearable biosensor technologies has transformed approaches to real-time health monitoring, yet challenges persist in achieving both mechanical robustness and reliable, continuous data analytics in dynamic environments. Conventional polymer-based sensing systems often fall short due to limited signal fidelity, inadequate adaptive analytics, or insufficient integration with secure, low-latency IoT frameworks. Addressing these deficiencies, this work introduces a flexible, deep learning-enhanced wearable biosensor platform that combines a nanostructured …
Published in Journal of Polymer & Composites · Vol. 13, Issue 6, 2025 · pp. 18–31 Read article
-
A Survey On Leveraging Machine Learning for Phishing Attack Prediction and Detection
Abstract: Phishing is one of the biggest cybersecurity threats that exploits user trust by masquerading as a legitimate site or email to steal personal and sensitive information. A state- of-the-art-phishing detection systems survey, this review showcases the evolution from traditional list-based techniques, including blacklisting and whitelisting to machine learning and deep learning models. While list-based systems cannot evolve to detect new and zero-day attacks, the ML algorithms of Decision Tree, Random …
Published in Journal of Microelectronics and Solid State Devices · Vol. 12, Issue 3, 2025 · pp. 1–10 Read article
-
Advanced Lithium-Ion Battery Prognostics: A Comprehensive Review of Machine Learning Approaches for Remaining Useful Life Prediction
Abstract: The lithium-ion battery (LIB), as one of the main sources for portable power systems, has been increasingly popular owing to its widespread applications in electric vehicles, consumer electronics, aerospace and renewable energy. Despite their advantages in high energy density and long cycle life, LIBs suffer from degradation over time of aging and cycling, resulting in loss of performance, safety issues, and economic bottlenecks. Predicting their Remaining Useful Life (RUL) is …
Published in International Journal of Mechanical Dynamics and Systems Analysis · Vol. 3, Issue 2, 2025 · pp. 12–27 Read article
-
Combining Unstructured and Structured Clinical Data in a Hybrid Transformer Model to Enhance Cardiovascular Analytics and Clinical Decision- Making
Abstract: Since cardiovascular disease (CVD) continues to be a major global cause of morbidity and mortality, early and accurate risk prediction is essential for prompt intervention and individualized treatment. This study introduces a new hybrid transformer-based model that combines unstructured clinical narratives, structured data, and customized lifestyle characteristics. A comprehensive understanding of disease progression is made possible by the model's ability to capture contextual, temporal, and patient- specific insights through the …
Published in International Journal of Bioinformatics and Computational Biology · Vol. 4, Issue 1, 2026 · pp. 30–37 Read article
-
Elderly Healthcare Using Federated Learning Approach
Abstract: The healthcare system for elderly people faces several challenges, which can be addressed using advanced machine learning models. These models can help monitor chronic diseases, detect falls, and provide personalized health recommendations. The study uses comprehensive datasets like MIMIC-III/IV, WESAD, and UCIHAR to explore human movements, device limitations, and the differences in fall occurrences. A detailed review of existing literature discusses current technologies for activity monitoring and fall detection, focusing …
Published in Current Trends in Information Technology · Vol. 16, Issue 1, 2025 · pp. 13–23 Read article
-
Leveraging Standards and Deep Learning Approaches to Secure Internet of Things (IoT) Devices from Cyber Attack
Abstract: The widespread adoption of Internet of Things (IoT) devices between 2019 and 2024 has significantly grows in various sectors in Japan, including healthcare, manufacturing, and the development of smart cities. Although this growth offers many advantages, it also makes these devices more vulnerable to cyber threats. High-profile security breaches in Japan have sparked discussions about the requirement for enhanced security measures to protect the rapidly evolving IoT technologies. This study …
Published in Journal of Artificial Intelligence Research & Advances · Vol. 13, Issue 1, 2025 Read article
-
Multimodal Data Fusion with Hybrid Machine Learning for Enhanced Prediction of Li-Ion Battery Remaining Useful Life and State of Charge
Abstract: Lithium-ion battery materials used in modern energy storage systems are required to exhibit high reliability, safety, and long lifecycle performance under varying operational and environmental conditions. Accurate prediction of Remaining Useful Life (RUL) and State of Charge (SoC) is therefore essential for understanding material degradation behavior, improving manufacturing quality, and enabling effective lifecycle management. However, nonlinear electrochemical aging, load variability, and thermal uncertainty significantly complicate accurate estimation of these parameters. …
Published in International Journal of Electro-Mechanics and Material Behaviour · Vol. 4, Issue 1, 2026 · pp. 1–5 Read article
-
Enhancing Maintenance Decision-Making in Thermal Power Plants Using Generative AI-Based Fault Diagnosis
Abstract: The growing complexity of operation and power consumption of thermal power stations involve the need to have intelligent fault diagnosis systems that can be used to guarantee reliability and safety in operation. In this research, a Generative AI (GenAI)-based hybrid architecture of early fault detection and predictive maintenance is proposed to improve the decision-making process of the maintenance team. The data-driven analytic approach combines methods of data-driven analytics, Generative AI …
Published in International Journal of Energy and Thermal Applications · Vol. 4, Issue 1, 2026 · pp. 25–33 Read article
-
Harnessing Deep Learning to Explore Microbial Community Structure and Carbon Storage Capacity in Mangrove Ecosystems: A Framework for Computationally
Abstract: Mangrove ecosystems represent one of the most efficient natural carbon sinks on Earth, functioning as critical blue carbon habitats that sustain diverse microbial communities responsible for biogeochemical cycling and long-term carbon storage. Despite their global ecological significance, accurately quantifying and predicting carbon sequestration in mangrove systems remains challenging due to the complex interactions between microbial diversity, sediment chemistry, and environmental drivers. This study presents a comprehensive and sustainable artificial intelligence …
Published in Research and Reviews : Journal of Computational Biology · Vol. 15, Issue 1, 2026 Read article
-
Utilizing Artificial Intelligence and Remote Sensing to Predict Flooding in Real-Time and Address Climate Resilience Policy in South Asia
Abstract: South Asia, a region characterized by hydro-climatic instability, faces an intensifying risk from devastating flooding, aggravated by human-induced climate change and intricate river basin interactions. Traditional flood prediction systems, based on limited in-situ data and resource-intensive physical models, have serious delays and resolution problems that make it harder to reduce disaster risk. The combined applications of Artificial Intelligence (AI) and high-resolution remote sensing (RS) constitute a paradigm shift in real-time …
Published in Journal of Remote Sensing & GIS · Vol. 17, Issue 1, 2026 Read article
-
Autonomous Agentic AI for Adaptive Cure Optimization and Defect Prevention in Thermoset Polymer Composite Manufacturing
Abstract: Thermoset polymer composites occupy a central position in modern structural manufacturing, from aircraft fuselages to wind-turbine blades. Despite progress in resin chemistry and fiber architecture, the “cure process” that transforms compliant preforms into load-bearing structures remains difficult to manage. Manufacturers encounter ‘voids’, “interlaminar delaminations”, and “spring-back distortion” when curing complex or thick-section parts. The cause is not ignorance of the relevant physics, but rather that ‘temperature’, ‘chemistry’, ‘rheology’, and ‘mechanics’ …
Published in Journal of Polymer & Composites · Vol. 14, Issue 2, 2026 · pp. 301–320 Read article
-
AI Driven IoT Based Decision Making System for Brain wave study: KSK approach for Brain wave study
Abstract: The rapid convergence of Internet of Things (IoT) architectures and deep learning has unlocked unprecedented potential for real-time neuro-diagnostic monitoring. This paper presents a novel framework for an AI-driven IoT ecosystem designed to capture, transmit, and interpret human electroencephalography (EEG) signals with minimal latency. Traditional brain-computer interface (BCI) studies are often constrained by localized computing power and the high dimensionality of neural data. Our proposed architecture integrates low-power EEG sensors …
Published in International Journal of Brain Sciences · Vol. 3, Issue 2, 2026 Read article
-
Learning Data Structures: Key to Good Programming
Abstract: Data structures are the most crucial feature of good programming and are needed to solve hard computational problems. This model makes use of two different recurrent neural network architectures, specifically long short-term memory (LSTM), and gated recurrent unit (GRU) networks. It explains how selecting and using the correct data structures may speed up computations, optimize memory, and scale code. How data structures and algorithms relate and how to think about …
Published in International Journal of Data Structure Studies · Vol. 4, Issue 1, 2026 · pp. 29–39 Read article
-
A Hybrid Multi-Physics Ensemble Deep Learning Framework for Simultaneous Prediction of Thermal and Electrical Conductivity in Functional Polymer-Based Nanocomposites
Abstract: Polymer-based nanocomposites have become promising materials for applications in energy storage, flexible electronics, biomedical devices, aerospace components, and advanced engineering because of their tunable thermal and electrical properties. Reliable prediction of these properties is essential for accelerating material design; however, existing analytical models and conventional machine learning techniques often fail to represent the complex interactions among filler characteristics, polymer matrices, processing conditions, and interfacial transport phenomena. This work presents HMEP-Net, …
Published in Journal of Polymer & Composites · Vol. 14, Issue 3, 2026 · pp. 1062–1082 Read article
-
A Review of AI-Based Intrusion Detection Systems for Mobile Ad Hoc Networks (MANETs)
Abstract: Mobile Ad Hoc Networks (MANETs) comprise wireless networks that lack any conventional infrastructure . Their chief features include highly changing network topologies, lack of centralized administration, and open nature of communication, which collectively result in making MANETs of the wireless kind very susceptible to a diverse range of cyber-attacks like blackhole, greyhole, wormhole, flooding, Sybil and denial-of-service (DoS) among others. Conventionally, Intrusion Detection Systems (IDS) relying on static rule-based methods …
Published in Journal of Artificial Intelligence Research & Advances · Vol. 13, Issue 3, 2025 Read article
-
Passive Digital Phenotyping for Longitudinal Burnout and Occupational Mental Health Surveillance: A Transformer-Based Explainable Deep Learning Approach Using Smartphone Behavioral Streams
Abstract: Occupational burnout constitutes a pervasive yet chronically under-surveilled public health threat, its insidious temporal evolution rendering episodic self-report instruments structurally inadequate for early detection. This paper introduces BurnoutSense, a passive digital phenotyping framework that continuously harvests eight heterogeneous smartphone behavioral data streams encompassing application usage ecology, communication metadata, geospatial mobility, screen interaction dynamics, inferred sleep rhythmicity, keystroke kinematics, ambient noise exposure, and battery/charging cadence to construct individualized multivariate behavioral signatures …
Published in International Journal of Behavioral Sciences · Vol. 3, Issue 2, 2026 · pp. 44–53 Read article