Browse Items (16481 total)
Sort by:
-
Predictive Modeling of Substance Abuse Risks using Big Data Analytics and Social Media Mining
The worldwide increase in substance abuse among teenagers and young adults has become serious concern in recent times. One way this pattern has developed is through the evolution of social media. Social media has transformed people's attitudes towards certain behaviors and has encouraged risky behavior to the point of actually causing addiction by exposing them to drug-related material. Despite the existence of preventative measures, such as education programs in schools, many children and youth have not had adequate access to educational interventions or evidence-based measures due to barriers created by geography, economic circumstances, and social factors, particularly in less developed countries. The research proposed is focusing on addressing this gap using a big data approach. This research employs a unique analytical framework that integrates multiple large data sets from a variety of sources to better identify and assess the effectiveness of interventions. This model employs an analytical approach that uses statistical learning techniques and predictive analytics to identify historical patterns and anticipate future trends, and assess the effectiveness of various interventions conducted in different countries. The results of the analysis suggest that this big data approach will provide decision-makers with clearly documented evidence related to various risk-taking behaviors as they relate to available prevention interventions, and will assist decision-makers in developing targeted prevention intervention strategies. This study demonstrates the revolutionary aspect behind the application of computational intelligence in preventing substance abuse and informing evidence-based community health interventions. 2025 IEEE. -
Hybrid Bidirectional GRU Approach for Crop Yield Prediction and Climate Change Impact Assessment in Agriculture
The impacts of climate change induced by humans will be felt most acutely by the agriculture sector due to its extreme dependence on weather. To ensure a steady supply of food, it is necessary to study and anticipate the effects of climate change on agricultural output. The impact of climate change on agricultural yield predictions is examined in this study using a novel methodology. In the proposed model, preprocessing, feature extraction, and training are the main processes. Data pretreatment guarantees quality by cleaning and normalising the data, while the PCC is utilised for feature selection. The model utilises AM and BiGRU for usage with large datasets. Using word vectors, the word embedding layer improves contextual awareness. Experiment findings show that the model is accurate to within 98.31% and can withstand a wide range of climate conditions. Current state-of-the-art methods are vastly outperformed by it, with performance measures like as R2 = 0.921%, MAE = 0.127%, and RMSE = 0.158%. These findings show that agricultural strategists and lawmakers can use AM-BiGRU to assess the effects of climate change and build a more resilient food system. 2025 IEEE. -
Automated Leaf Disease Detection using a Hybrid CNN-BiLSTM Model for Smart Agriculture
The mitigation of crop losses and the sustainability of agriculture rely on the prompt identification of foliar diseases. In large-scale agriculture, conventional identification methods such as expert eye inspections are inefficient, susceptible to errors, and labour-intensive. A growing number of individuals are seeking automated methods to monitor plant health, given that the majority of Indians are employed in agriculture. This study presents a hybrid DL strategy for leaf disease detection, encompassing preprocessing, segmentation, feature extraction, and model training. Initially, images are processed to enhance their quality and uniformity. The impacted regions of the leaf are subsequently categorised by K-Means clustering. The classification accuracy is improved by utilising several feature extraction methods. The proposed model, CNBiLS, integrates bidirectional LSTM layers with convolutional layers to leverage the spatial and sequential information in image data. When evaluated against contemporary state-of-the-art models, CNBiLS exhibited superior performance, achieving an exceptional 99.84% classification accuracy. This result underscores the model's accuracy in identifying various leaf diseases. Ultimately, CNBiLS offers a precise, scalable, and robust automated system for detecting leaf diseases, equipping farmers with timely information to manage illnesses effectively, so enhancing both the quality and yield of their crops. 2025 IEEE. -
Deep Q-Learning for Autonomous Vehicle Navigation in Smart Mobility
The proposed system leverages Deep Q-Learning to enhance autonomous vehicle navigation in smart mobility environments. By integrating reinforcement learning with deep neural networks, the system enables vehicles to make real-time decisions while adapting to dynamic traffic conditions. The framework employs a reward-based learning mechanism to optimize path selection, collision avoidance, and efficient maneuvering in complex urban scenarios. To improve decisionmaking accuracy, the proposed approach incorporates an experience replay mechanism, preventing overfitting and ensuring stable learning. Additionally, a target network is utilized to enhance training convergence, allowing the model to generalize effectively across varying road conditions. The system is further optimized through adaptive explorationexploitation strategies, enabling vehicles to balance learning new routes while prioritizing safe and efficient navigation. The proposed methodology demonstrates significant improvements in autonomous mobility, offering a scalable and robust solution for next-generation smart transportation systems. 2025 IEEE. -
Integrating Explainability and Fairness in Credit Risk Prediction: A Hybrid Approach Using Tabnet, LightGBM, SHAP, and Counterfactual Explanations on the FICO Dataset
Transparent and fair credit risk assessment is essential for responsible lending in modern financial systems. This paper presents an interpretable and ethically grounded machine learning framework for loan default prediction using the FICO Explainability Challenge dataset. We combine LightGBM, a high-performing gradient boosting model for tabular data, with TabNet, a deep learning architecture that provides intrinsic interpretability through attentive feature selection. To enhance transparency, SHapley Additive exPlanations (SHAP) are employed for global and local feature attribution, while counterfactual explanations generated using the DiCE framework offer actionable recourse. Fairness is evaluated and mitigated using IBM's AI Fairness 360 toolkit. Experimental results demonstrate that the proposed hybrid approach achieves strong predictive performance while ensuring interpretability and fairness, making it suitable for trustworthy and regulation-compliant credit risk modeling. 2026 IEEE. -
Quantum-Enhanced Cryptographic Key Exchange for Secure IoT Networks
The coupling of quantum-metering methods with cryptographic key exchanges spells out a new security paradigm model to safeguard Internet of Things (IoT) networks against rapidly changing cyberspace threats. The given assessment explains a quantum-advanced scenario that is operationally feasible and serves as a platform for quantum key distribution (QKD) along with classical post-quantum algorithms to deliver end-to-end confidentiality, integrity, and authentication features to a wide range of IoT devices. The model, by the virtue of quantum entanglement and photon polarization used in the creation of tamper-evident communication links, is invulnerable to adversarial-aided eavesdropping and computational assault methods and further offers a hybrid encryption protocol that has been demonstrated to alleviate key generation and exchange latency trade-offs - simultaneously maintaining scalability and effectiveness in resource-constrained IoT node environments. The key compromise probability is significantly lowered in the experimental results, along with the keys' entropy levels, which were generated in the pure QKD space as opposed to classicalbased RSA and ECC methods. Also, the framework investigates lightweight quantum-safe authentication techniques that can be used to establish trust at the device level. The output points to the enhanced resistance to quantum and classical attackers that can make the solutions work in real-time application IoT environments such as smart healthcare, autonomous systems, and industrial automation. Overall, the quantum-enriched model of cryptographic key exchange is next-gen IoT ecosystem to be implemented subsequently. 2026 IEEE. -
Comparative analysis on containers based on kubernetes, Docker swarm, open shift and Mesos
Cloud Computing has become increasingly common to use cloud computing solutions for Big Data processing because of their vast variety of computer resources and ability to extend over multiple cloud platforms. The rapid growth of the Internet of Things (IoT) concept has sparked this development. A traditional approach of cloud organisation is virtualization, which uses virtual computers and containers. It is impossible to overestimate the importance of lightweight cloud infrastructure for microservices. Many academics have proposed container-based virtualized computing services as a result of this. Container technology has risen in prominence as a viable alternative to traditional virtual machines in recent years. There is need to exploit high-level services such as orchestration. In this paper, we compare performance of various container orchestrators like Kubernetes, Dockswarm, Openshift and Mesos. 2025 IEEE. -
Modern Approaches for Automatic Question Paper Generator
The Automatic question paper generation system in the educational field can be useful in improving the quality of question paper, distribution and evaluation process. The system can be used for maintaining the quality of the questions with higher accuracy and less error rate compared with other existing systems at a lower cost. This article gives a comprehensive literature survey of modern approaches followed in automatic question generation (AQPG) systems and categorizing the approaches used for the question paper generation process. The techniques such as rule-based, encoder-decoder based, generative adversarial network (GAN)-based, reinforcement learning-based, and transformer-based approaches are discussed in the paper and evaluated using standard metrics. The article presents insights into the strengths and limitations of each approach through the systematic comparison and analysis of multiple studies using BLEU-4, ROUGE-L, and METEOR metrics on the SQuAD dataset. The research finding of the article gives a better opportunity to the researcher and educators to improve the knowledge about automated question paper generator systems as well as the challenges incorporated during the implementation process of question paper generation. This article also gives a depiction of AI enabled solution in automated question paper generator. 2025 IEEE. -
Cost Sensitive based Support Vector Machine for Energy Efficient Clustering and Routing
Wireless Sensor Networks (WSNs) are ideal for applications requiring rapid deployment, as they operate without the need for a pre-existing network infrastructure. WSNs have constrained capacity to manage large volumes of data, and processing, transmitting as well as receiving this data consumes substantial energy. On account of their constrained power usage and bandwidth, sensors are unable to transmit each gathered information to a Base Station (BS) or sink for analysis. Thus, this research proposes the Cost Sensitive based Support Vector Machine (CS-SVM) approach for the energy efficient clustering and routing. The proposed approach minimizes an energy consumption at data transmission and clustering, extending network lifetime through effectively selects CHs according to energy levels as well as node distribution. The CH selection considers the three important fitness functions such as distance from sink to CH, distance from CH to sensor nodes and residual energy. Then, the routing considers the two important fitness functions. The various performance indices are considered to validate the effectiveness of proposed method. The simulation outcomes proven that the proposed CS-SVM approach attains the better energy consumption of 8.32J as compared to the previous approach named Distributed Energy-efficient two-hop-based Clustering and Routing (DECR). 2025 IEEE. -
The Role of Machine Learning in Shaping Virtual Reality Educational Experiences: A Multi-Model Analysis
The integration of Virtual Reality (VR) in educational settings offers unique opportunities for enhancing learning experiences and outcomes. This study evaluates the efficacy of various machine learning models in predicting student engagement and educational outcomes within VR-enhanced learning environments. Utilizing a dataset comprising 5,000 entries related to VR usage in education, we applied both classification and regression machine learning techniques to predict binary outcomes (e.g., the usage of VR in education) and continuous outcomes (e.g., levels of student engagement). Models such as Logistic Regression, Random Forest, Support Vector Machines (SVM), K-Nearest Neighbors (KNN), Decision Trees, and advanced regression techniques including Ridge, Lasso, Polynomial Regression, and Support Vector Regression (SVR) were systematically analyzed. The performance of each model was assessed based on accuracy, precision, recall, F1-score, R2, MSE, and MAE. We find that models such as SVM and Random Forest performed well for classification tasks and handled imbalances in classes gracefully, while SVR and Random Forest Regressor did a better job at regression tasks, being able to capture complex, nonlinear relationships in the data. It highlights the possibility that machine learning can accurately predict outcomes of VR engagement and perhaps can help inform VR-based education design to be more effective. The goal of this comparative analysis is to provide guidance for educators and technologists in deciding which machine learning strategy is suitable to facilitate education in VR with regard to improving educational outcomes. 2025 IEEE. -
License Plate Recognition Model based on Improved YOLOv5 and Convolutional LSTM
An end-to-end deep learning model is proposed in this research, for licence plate recognition (LPR) and identification in natural circumstances, which addresses the accuracy and speed limitations of standard licence plate recognition approaches. By adding a better channel attention mechanism and including position data in the output, the proposed method improves the You Only Look Once (YOLOv5) down sampling process and reduces information loss during sampling for better feature extraction. An optimised the YOLO layer is used for single-class recognition to improve efficiency and accuracy. Additionally, Convolutional Long Short-Term Memory (ConvLSTM) combined with Connectionist Temporal Softmax (CTS) is used for character segmentation-free recognition. The utilization of an optimized YOLO layer for single-class recognition enhances both efficiency and accuracy. The integration of ConvLSTM in conjunction with CTS proves to be a breakthrough, facilitating faster convergence, reduced training time, and increase the precision of the model. This configuration speeds up convergence, lowers training time, and increases identification accuracy. The experimental results demonstrate average recognition precision of 99.24% and also robustness, especially in complex situations, with better performance than conventional algorithms. 2025 IEEE. -
Multilingual Sentiment Analytics for India's NEP 2020
This study presents a multilingual sentiment analysis framework for evaluating public sentiments on India's National Education Policy (NEP) 2020. The authors developed a dataset related to NEP 2020 using web scraping from open sources. The curated dataset comprises 50,000 social media posts (English: 30,000, Hindi: 12,000, Tamil: 8,000) processed through a confidence-gated hybrid annotation pipeline. Sentiment labels were created using Transformer models (BERT, mBERT, XLMR) and validated by native-speaker with F1-scores of 87.6%, 81.2% and 78.0% for English, Hindi and Tamil respectively: outperforming baselines (SVM, Naive Bayes, BiLSTM) by 12-18% (p<0.001). We use computational efficiency measures to illustrate that training takes 3.2-5.3 hours and inference lasts between 118 and 187 posts per second. Topic modeling revealed sentiment divergences: positive for linguistic inclusivity and teacher training, negative for affordability and infrastructure. Cross-linguistic analysis showed English-Hindi convergence (similarity: 0.61) versus Tamil divergence (0.46), reflecting regional priorities. Tamil emphasized linguistic identity while English prioritized implementation critiques. Quantitative policy impact analysis shows very strong correlation (r=0.68, p<0.01) between regional sentiment scores and state adoption rates. This open-sourced contribution is filling the crucial gap of inclusive policy analytics in multilingual society informing evidence-based policy. 2025 IEEE. -
Advanced Malware Analysis and Detection Using Deep Neural Networks
Malware is malicious software that is used to cause harm to the computer systems, networks or users across several operating systems, such as Windows, macOS, iOS, Android and Linux. The identification and categorisation of malware is a difficult subject with no one-size-fits-all solution due to the constant evolution of the malware and lack of standardised detection frameworks. The use of deep learning models in cybersecurity encourages the growth of Explainable Artificial Intelligence (XAI) and Interpretable Machine Learning (IML) techniques. The goal of this research is to automate the way of analysing malware using a simple framework without the need of complex software. This article focuses on the application of Neural Networks, a deep learning model in detecting and analysing the behavior of malware. The model performed better when compared to other techniques by achieving an accuracy of 99.43%, precision of 99.05% and a F1 score of 0.99, which was trained on a large dataset containing 1,38,048 samples. 2025 IEEE. -
Advanced Malware Analysis and Detection Using Deep Neural Networks
Malware is malicious software that is used to cause harm to the computer systems, networks or users across several operating systems, such as Windows, macOS, iOS, Android and Linux. The identification and categorisation of malware is a difficult subject with no one-size-fits-all solution due to the constant evolution of the malware and lack of standardised detection frameworks. The use of deep learning models in cybersecurity encourages the growth of Explainable Artificial Intelligence (XAI) and Interpretable Machine Learning (IML) techniques. The goal of this research is to automate the way of analysing malware using a simple framework without the need of complex software. This article focuses on the application of Neural Networks, a deep learning model in detecting and analysing the behavior of malware. The model performed better when compared to other techniques by achieving an accuracy of 99.43%, precision of 99.05% and a F1 score of 0.99, which was trained on a large dataset containing 1,38,048 samples. 2025 IEEE. -
Interpretable Deep Learning for Multiclass Psychological Disorder Classification Using CNN and TCAV Using fMRI
In this paper a 3D convolutional neural network model is presented that has been improved with squeeze-and-excitation blocks and residual blocks to categorize functional MRI data across healthy controls, schizophrenia, and attention deficit hyperactivity disorder classes. A post hoc explainability framework was applied to concept activation vectors based on mean activation for each class and testing using TCAV. The merging of CAV and TCAV for fMRI data was to improve transparency by providing an interpretable model. This helps in understanding the prediction sensitivity of the models. Concept vectors are defined through the extraction and analysis of intermediate activations from the CNN-SE model. These vectors are then used to calculate the TCAV scores, which indicate the degree to which each brain region influences the model output. The precision of the deep learning model was up to 79%. Similarity matrices indicate the degree of overlap and correlate with the model's results. Visualizations such as activation heat maps and glass brain overlays based on the AAL atlas further support the interpretability of the model, making it more transparent and suitable for clinical applications. Variations in activation between classes can be observed using a visual feature plot. This framework allows mapping model predictions to interpretable neuroanatomical regions and identifies classspecific dependencies on particular brain networks. 2025 IEEE. -
Patient Digital Twins for Dynamic Hospital Supply Chain Management AI-Based Predictive Resource Allocation
The fusion of Digital Twin (DT) technology and Artificial Intelligence (AI) holds great promise for the optimization of hospital supply chain management and resource allocation. This paper proposes a patient-specific digital twin framework aimed at forecasting hospital staffing needs and aiding supply planning by means of AI-based analytics. Preprocessing and modeling of hospital records that included admission information, medical procedures, room types, usage of supplies, and patterns of staffing were achieved by utilizing advanced machine learning algorithms. The strategy illustrated better performance compared to baseline strategies and had interpretability from feature importance analysis, which emphasized length of stay, critical care admissions, and specialized procedures as influential drivers of staffing requirements. The results show that the patient digital twin can optimize operational efficiency, avoid supply deficits, and facilitate evidence-based decisions. The suggested framework is consistent with the vision of contemporary healthcare supply chains in that it promotes resilience, flexibility, and smart resource management. 2025 IEEE. -
Data-Driven Transformation of Hospitality Supply Chains Using AI-Powered Segmentation
The increasing complexity of supply chain operations in the hospitality sector demands data-driven strategies for efficient resource utilization and service delivery. This study proposes an artificial intelligence (AI)-driven framework leveraging unsupervised machine learning to uncover hidden patterns in patient-related operational data sourced from a publicly available dataset. The research applies clustering algorithmsK-Means, DBSCAN, and Agglomerative Hierarchical Clusteringto segment patient prof iles based on key variables such as length of stay, procedure type, room category, equipment usage, and staffing needs. Principal Component Analysis (PCA) was employed for dimensionality reduction and cluster visualization. The optimal number of clusters was identified using the Elbow Method, with K-Means yielding the highest silhouette score. Comparative analysis of all clustering models revealed varying strengths in noise detection, interpretability, and handling of sparse features. The results demonstrate how intelligent segmentation can support dynamic resource planning, targeted supply allocation, and improved operational responsiveness in hospital-based hospitality systems. This work contributes to the growing domain of AI-enabled supply chain analytics and of fers a practical pathway for enhancing decision-making in smart hospitality environments. 2025 IEEE. -
A Lightweight Hybrid Deep Learning model for Real-Time Intrusion Detection in IoT Networks.
In IoT networks, cyber threats are hard to identify because of dynamic and heterogenic nature of IoT traffic as it cannot be identified with more traditional intrusion detection systems. This paper discusses the deep learning methods of intrusion detection in terms of CNN+LSTM, CNN+BiLSTM, CNN+GRU, and BiGRU+RF and propose a novel Dense+SimpleRNN architecture. Preprocessing includes label encoding, feature selection, normalization, SMOTE balancing, and reshaping sequences, using the RT-IoT 2022 dataset. The paper demonstrates, CNN + BiLSTM and CNN + GRU achieving similar accuracy but with higher computational cost. On the other hand, the proposed Dense+SimpleRNN has 98.59% accuracy, precision and recall and Fl-score, which are higher than the baselines models. The results point that Dense+SimpleRNN is an efficient and lightweight IDS that is very appropriate in real-time IoT network security. 2025 IEEE. -
A Hybrid Stacked Ensemble Model for Heart Disease Prediction
Cardiovascular Diseases (CVDs), especially heart attacks, are resulting in high rates of death worldwide, which highlights the need for early prediction systems. This paper deals with advanced ML and DL methods to predict heart attacks with a pre-processed clinical dataset. 6 models were used: a Hybrid Stacked Model combined with Logistic Regression, Random Forest, and XGBoost using a neural meta-learner; CNN with LSTM, BiGRU, and dense layers; an RNN with BiLSTM; and an XGBoost method using deep feature representations. Data preprocessing involved feature scaling and class balancing with the help of SMOTE. Model performance is being measured by Accuracy, Precision, Recall, and F1-Score. Hybrid Stacked Model had the highest accuracy (94.24%) and F1-score (94.12%), while CNN + LSTM had the best recall (95.96%), to reduce false negatives. XGBoost with deep features demonstrated competitive accuracy (91.22%) and transparency. These results point to the efficiency of hybrid and sequential deep learning models in cardiovascular risk prediction. In the future, research will be focused on real-time patient data integration, federated learning for privacy, and personalized health promotion using IoT-based monitoring. 2025 IEEE. -
Explainable AI for Heart Disease prediction: A Clinical Transparency Route Experiment
In this paper, a proposeable explainable machine learning procedure on estimating the danger of heart attack will be proposed with a stacked ensemble of XGBoost, Random Forest, and Multi-layered perceptron (MLP). The data set of UCI Heart Disease was preprocessed by normalization, imputation, and SMOTE to address the imbalance problem and the variables were optimized with the help of the feature engineering. The model performance was measured using accuracy, precision, recall, F1-score and ROC-AUC. In order to make the results more interpretable, Explainable AI were applied with SHAP and LIME, and the most relevant risk factors including troponin, cholesterol, and blood pressure were indicated.. In this paper, it is shown that ensemble learning in XAI can yield plausible, interpretable, and clinically practical data to complement enhanced cardiovascular diagnostics. 2025 IEEE.