Search

The exponential growth of intelligent systems technologies, including Artificial Intelligence (AI), Internet of Things (IoT), Machine Learning (ML), and Smart Connected Products, has intensified the difficulties of data governance. Organizations adopting these technologies face challenges in managing the volume, variety, and velocity of data while ensuring quality, security, compliance, and ethical integrity. This study investigates the state of big data governance (BDG) programs in organizations leveraging intelligent systems technologies, using a systematic literature review guided by PRISMA. A synthesis of insights from 74 peer-reviewed articles, conferences, and industry reports reveals both the significant benefits of BDG such as enhanced decision-making, operational efficiency, and regulatory compliance and persistent challenges, including fragmented governance frameworks, limited scalability, data quality issues, and ethical concerns. To address these gaps, this study proposes the big data governance maturity assessment model (BDG MAM), a novel framework developed to assess and enhance the maturity of BDG programs. The BDG MAM evaluates governance maturity across four key dimensions: people, process, data, and technology. It provides a structured roadmap for organizations to benchmark governance practices, prioritize improvements, and implement effective strategies. The model was validated through a pilot study conducted with a public higher learning institution, demonstrating its practical applicability and effectiveness in real-world scenarios. By bridging theoretical insights with practical implementation, this study advances academic discourse and provides practitioners with a robust approach to navigate the challenges of modern data governance practices, ensuring sustainable and effective management of heterogeneous data environment.

In the remote healthcare industry data analytics denotes the computerization of collection, processing, and exploring complicated data to acquire finer perceptions and validate healthcare practitioners to produce familiar decisions. Healthcare basics in the modern age are vital challenges specifically in developing countries owing to the shortfall of difficult hospitals and medical professionals. As fuzzy systems have reformed several areas of work, health has also made the most of it. In this paper, the purpose of the study is to introduce a novel and intelligent remote healthcare system based on modern technologies like the Internet of things (IoT) and Neutrosophic fuzzy systems to ensure precise data analysis with lesser time and energy consumption. In this study, a novel method called, Blinder Oaxaca-based Shapiro Wilk Neutrosophic Fuzzy (BO-SWNF) data analytics for remote healthcare is designed. Data collection is performed with the WESAD dataset. Duplicated data are eliminated by Blinder Oaxaca Linear Regression-based Preprocessing model. With the application of the Blinder Oaxaca function, energy efficiency is enhanced. Finally, the Shapiro Wilk Neutrosophic Fuzzy algorithm is applied for ensuring robust data analysis. The experimental results of the proposed BO-SWNF envisage the data for finer comprehension of attribute distribution. The result is conducted by using PYTHON application to analyze stress detection with the WESAD dataset. The proposed BO-SWNF method achieved an overall accurate data analysis of 12 % with minimum time ensuring 56 % improvement and minimizing energy consumption by 54 %.

The emergence of large language models (LLMs) has revolutionized various fields, including education, finance, marketing, healthcare, and medicine. In this review, we aim to explore the application of LLMs in the healthcare sector, with a specific focus on disease diagnostics. The review highlighted the widespread use of LLMs, such as GPT-4, ChatGPT, GPT-3.5, and LLaMA, with GPT-4 being the most frequently used in disease diagnostics due to its diverse applications, improved accuracy, and efficiency. This review shows that LLMs have utilized a variety of medical data sources, including general medical databases, specialized documents, medical images, and genomic data. Moreover, the focus of these LLMs spans a broad spectrum of healthcare fields, addressing chronic conditions, respiratory diseases, cancer, and rare diseases. The performance evaluation of LLMs involves both qualitative and quantitative measures assessing their diagnostic accuracy. The findings highlight the evolving nature of LLMs in improving diagnostic accuracy.

Early diagnosis and effective treatment of age-related macular degeneration (AMD), a leading cause of vision impairment, are critically dependent on accurate grading. This paper presents a novel framework, named Mask-UnMask Regions (MUMR), designed to differentiate between normal retina, intermediate AMD, geographic atrophy (GA), and wet AMD using standardized retinal fundus images with an input resolution of $1024 \times 1024$ pixels. The framework initiates with the downscaling of images to a quarter of their original size via a Preserving High-Frequency Information (PHFI) module, which retains key details essential for further analysis. Additionally, we developed a simple, lightweight, yet efficient ResNet-like network for feature extraction and introduced a Region Interaction (RI) module. This module incorporates Adaptive Mask and UnMask Sub-Modules, identifying significant regions while reconstructing less relevant areas using a direction-constrained self-attention mechanism to ensure the learning of global structural cues critical for AMD grade classification. The proposed method was evaluated on a dataset comprising 864 retinal fundus images. Our model consistently outperforms state-of-the-art approaches, achieving mean accuracy, mean F1-score, and mean Cohen’s Kappa of 92.55%, 92.59%, and 89.97%, respectively. In the binary classification task of distinguishing between Non-AMD and AMD cases, the proposed approach also surpasses competing models, achieving mean accuracy, mean F1-score, and mean Cohen’s Kappa of 97.11%, 97.03%, and 94.06%, respectively. Furthermore, statistical analysis of these metrics confirms that the improvements are statistically significant, demonstrating the robustness and improved performance of our proposed framework in AMD grading.

Modulated Model Predictive Control (MMPC) techniques can be applied to enhance matrix converter (MC) performance in the linear modulation zone (LMZ). However, extending the converter operation in the overmodulation zone (OMZ) remains a problem. A key limitation of existing control methods in the OMZ is using the duty cycles calculated for the LMZ without considering the voltage or current reference vectors. This approach can lead to inaccurate calculations and a slower transient response when switching between the two operational zones. To overcome this challenge, a novel reference vector calculation method is proposed in this paper. The proposed approach simplifies the duty cycle calculation and ensures optimal performance in the OMZ by introducing new vector calculations and dividing the OMZ into two subzones. This method also minimizes the transition time between LMZ and OMZs to maximize the usable operating range of MCs. Simulation and experimental results validate the proposed method outperforms two other MMPC methods in improving the MC performance, reducing the transition times between operational zones, and maximizing the converter utilization.

Cancer treatment is a challenging endeavor because of the intricacy, heterogeneity, and diversity of cancer causes. Comprehensive therapeutic approaches are crucial for cancer treatment. Anticancer peptides (ACPs) present a potentially effective therapeutic option. However, the extensive identification and synthesis of these peptides present a persistent difficulty that calls for the creation of effective prediction techniques. Existing techniques either suffer from low accuracy or employ high-dimensional feature sets, frequently producing sparse features and leading to ineffective model designs. This work presents a novel set of features and a long-short-term-memory (LSTM)-based classification strategy to create an efficient model. The suggested feature set includes three new and two modern feature extraction methods. The binary profile feature and k-mer sparse matrix of the reduced amino acid alphabet are part of the modern feature set. The combination of the composition of the K-spaced side chain pairs (CKSSCP), the composition of the K-spaced electrically charged side chain pairs (CKSECSCP), and the combination of [pk(CO2H)] + [pk(NH2)] + [pk(R)] + [isoelectric point] is used to derive the novel features. The suggested LSTM model is trained using the combined feature set. The trials are carried out with a k-fold cross-validation method on benchmark datasets. The results indicate that the proposed model outperforms alternative ACP classification techniques in terms of Mathew’s correlation coefficient (MCC) and accuracy. The ACP740 dataset with 5-folds yields an MCC score of 75%, which is 12%, 11%, 3%, and 8% greater than those of the ACP-DL, ACP-DA, ACP-MHCNN, and ACP-KSRC approaches, respectively. For the ACP344 dataset with 10-folds, the proposed method achieves an MCC score of 85.14%, which is 23% and 2% higher than the MCC scores of ACP-DL and SAP methods, respectively. Better classification performance offered by the proposed approach could help identify new ACPs and better understand their structural and chemical characteristics. The source code and the datasets are available on the author’s GitHub page (https://github.com/Shujaat123/ACP-LSTM-NFR).

The growing deployment of Internet of Things (IoT) devices in diverse daily-life smart applications makes the security of IoT networks a critical requirement. However, efficient support of network security remains challenging even for a standardized IoT network protocol such as the IPv6 Routing Protocol for Low Power and Lossy Networks (RPL). It incorporates limited protection from external security attacks but stays considerably vulnerable to internal routing attacks. The inherent design of RPL, particularly its topology establishment and maintenance mechanism, makes it easy to initiate more damaging attacks such as flooding attacks. Given the constrained characteristics of IoT devices, flooding IoT networks can easily lead to resource exhaustion and network performance degradation. This paper introduces a new variant of the flooding attack namely the Destination Advertisement Object Flooding (DAOF) attack. It is based on using routing communications, particularly the DAO messages, to disseminate unnecessary routing traffic which instigate excessive transmissions of control messages across the network. As demonstrated by the experimental results of this study, the DAOF attack can increase network overhead by more than 65% even in a relatively small-scale setup. Additionally, it can notably lead to high energy consumption with an average increase of 36% and low QoS performance with an average latency increase of 150%. For effective mitigation of the DAOF attack, a new lightweight solution based on a simple collaboration among RPL nodes is presented in this paper. It is referred to as DAOF-Secure RPL (DAOF-SRPL). It is based on simple in-protocol modifications to provide RPL with effective security support against the DAOF attack. In contrast to RPL in the attack scenarios, DAOF-SRPL achieved a reduction of over 80% in the total transmission rates of control messages. Meanwhile, it was able to maintain energy consumption and latency at minimal levels while preserving the same PDR results.

The electrification of aviation is predominantly driven by environmental concerns associated with emissions. In this context, while the More Electric Aircraft and Hybrid Electric Aircraft concepts are currently at the forefront, the long-term objective is the utilization of All Electric Aircraft. Presently, the transition to All Electric Aircraft, which includes the electrified propulsion system, seems challenging due to the significant energy and power demands of aircrafts, given that the specific energy of recent energy storage technologies and power density of other electrical parts are not at the desired levels. Nevertheless, various scientific findings are being explored to accelerate this process and fuel cells are emerging as a solution. This study examines the retrofitting of a Cessna 172S, selected as the baseline aircraft, into a fully electric version with a hybrid configuration combining a battery and fuel cell. The modification process includes the development of an electrical power system architecture, selection of the required components, and modeling of the system elements. The new configuration of the platform is investigated through various simulations to evaluate the impacts on power sharing of the battery and fuel cell, and a comparative weight analysis is conducted between the baseline and retrofitted aircraft. The study concludes with a critical examination of the benefits and challenges related to the All Electric Aircraft concept.

New technologies drive educational shifts, transforming offline to online learning. This study investigates e-learning and cloud computing integration to understand synergies and their potential impact. The study addresses two primary research questions: the influence of e-learning on factors like architecture, software, performance, security, hardware, network, and virtual aspects, and the examination of cloud computing services and models such as SaaS, PaaS, IaaS, and S.O.A. The research aims to provide insights into how e-learning is incorporated in a cloud computing environment. The motivation behind this study is to investigate the intricate relationship between e-learning and cloud computing. By analyzing 154 scientific papers, the study delves into the specifics of this integration, highlighting trends and areas that have received more attention. The study examines e-learning in a cloud computing environment, focusing on architecture (27%), general topics (21%), software (19%), and performance (18%). Virtual environments have fewer security issues, while storage and network focus are more prevalent. Cloud computing services are mainly all services, with software as a service (18%), infrastructure as a service (17%), and platforms as a service (10%). Most studies are based on public clouds (74%), all other models (11%), and hybrid clouds (3%). The study examines e-learning integration in cloud computing, highlighting limitations in hybrid and private clouds, specialized infrastructure, and a gap in platforms and infrastructure offerings.

The emergence of the sixth generation (6G) era has highlighted the importance of Network Slicing (NS) technology as a promising solution for catering the diverse service requests of users. With the presence of a large number of devices with different service requests and since each service has different goals and requirements; efficiently allocating Resource Blocks (RBs) to each network slice is a challenging task to meet the desired Quality of Service (QoS) standards. However, it is worth noting that the majority of research efforts have primarily concentrated on cellular technologies, leaving behind the potential benefits of utilizing unlicensed bands to alleviate traffic congestion and enhance the capacity of existing LTE networks. In this paper we propose a novel idea by exploiting LTE-WLAN Aggregation technology (LWA) in Multi-Radio Access Technology (RAT) Heterogeneous Networks (HetNet), aiming to solve radio resource allocation problem based on the Radio Access Network (RAN) slicing and 5G New Radio (NR) scalable numerology technique. A joint optimization problem is proposed by jointly finding an efficient resource allocation ratio for each slice in each Base Station (BS) and by finding the optimum value of scalable numerology with the objective of maximizing users’ satisfaction. In order to solve this problem, a novel three-stage framework is proposed which is based on channel state information as a pre-association stage, Reinforcement Learning (RL) algorithm as finding the optimum value of slice resource ratio and scalable numerology, and finally Regret Learning Algorithm (RLA) as users’ re-association phase. Furthermore, a comprehensive performance evaluation is conducted against different baseline approaches. The simulation results show that our proposed model balances and achieves improvement in users’ satisfaction by deploying the proposed Multi-RAT Het-Net architecture that leverages LWA technology.

Practical experiments are a crucial step to demonstrate the viability of reconfigurable intelligent surface (RIS)-empowered communication, which is one of the emerging technologies for next-generation networks. In this paper, we present practical measurements to demonstrate the RIS capabilities for enhancing signal coverage and providing physical layer security (PLS) in an indoor environment. First, extensive measurements are performed in a single-user deployment using iterative, grouping, and codebook-based phase shift optimization methods. The iterative method achieves approximately 10 dB performance improvement in the received signal power through careful adjustments of RIS phase configurations when the receiver is placed at different locations. The grouping method reduces the training time to find a suitable RIS configuration by sacrificing only a few dBs in the received signal power. Another set of experiments is conducted for a multi-user deployment to exhibit PLS, where the RIS is partitioned into two regions serving the intended and unintended users. The results demonstrate that the codebook method can effectively boost the secrecy capacity on the move without utilizing feedback other than the users’ positions during the phase shift optimization process, while the iterative method requires a continuous feedback channel for the received signal powers.

Backpropagation neural networks are commonly utilized to solve complicated issues in various disciplines. However, optimizing their settings remains a significant task. Traditional gradient-based optimization methods, such as stochastic gradient descent (SGD), often exhibit slow convergence and hyperparameter sensitivity. An adaptive stochastic conjugate gradient (ASCG) optimization strategy for backpropagation neural networks is proposed in this study. ASCG combines the advantages of stochastic optimization and conjugate gradient techniques to increase training efficiency and convergence speed. Based on the observed gradients, the algorithm adaptively calculates the learning rate and search direction at each iteration, allowing for quicker convergence and greater generalization. Experimental findings on benchmark datasets show that ASCG optimization outperforms standard optimization techniques regarding convergence time and model performance. The proposed ASCG algorithm provides a viable method for improving the training process of backpropagation neural networks, making them more successful in tackling complicated problems across several domains. As a result, the information for initial seeds formed while the model is being trained grows. The coordinated efforts of ASCG’s Conjugate Gradient and ASCG components improve learning and achieve global minima. Our results indicate that our ASCG algorithm achieves 21 percent higher accuracy on the HMT dataset and performs better than existing methods on other datasets(DIR-Lab dataset). The experimentation revealed that the conjugate gradient has an efficiency of 95 percent when utilizing the principal component analysis features, compared to 94 percent when using the correlation heatmap features selection approach with MSE of 0.0678.

Recently, the huge growth in multimedia demand derived from smart devices and Internet of Things (IoT) development requires efficient heterogeneous networks with high data rates. To fulfill these requirements, the exploitation of Unmanned Aerial Vehicles (UAVs) enabled with multiple radio access for providing efficient connectivity should be considered, since it offers flexibility, ease of deployment, and on-demand service for wireless devices. In this paper, the idea of deploying Multiple Radio Access Technologies (Multi-RAT) base stations on a UAV is proposed, in order to utilize the unlicensed spectrum. The problem of optimizing the Multi-RAT UAV’s location along with the wireless devices’ association to maximize the total system throughput considering a heterogeneous LTE and WLAN ground network is investigated. To solve this problem, we propose a novel framework based on reinforcement learning and regret matching algorithms, such that the Q-learning algorithm is used to find the Multi-RAT UAV’s optimum location, while regret matching is used to solve the optimum users’ association. Moreover, a K-means clustering algorithm is adopted as an initialization phase to speed up the convergence of the proposed solution. Simulation results show the significance of the proposed idea of deploying Multi-RAT base stations on a UAV. Further, the performance analysis shows the effectiveness of the proposed framework by applying K-means as an initialization phase.

This study proposes a novel isolated multiport bidirectional DC/DC converter (IMBC) that combines soft-switching (SS) and resistor-capacitor-diode (RCD) snubber circuits for battery charge and discharge applications in photovoltaic (PV) power systems. The proposed SS circuit operates at zero-current (ZC) switching (ZCS) when the switches are on to reduce the switching losses. The RCD snubber circuit also suppresses the overvoltage when turning off switches. Combining the two circuits also enables a more comprehensive operating range for SS, resulting in low losses and high efficiency. A thorough study of operating systems and design considerations is shown. PSIM illustrates the effectiveness of the suggested control method. A prototype of the suggested IMBC with an output voltage of 120 V and an output power of 150 W is constructed to confirm the viability and efficiency of the suggested topology. The maximum efficacy of the suggested topology in the step-up mode for the PV arrays is about 96.4% at an input voltage of 20.5 V and output power of 150 W. In comparison, the efficiency is about 98.5% in the step-down mode for charging and discharging the battery which is considered a highly satisfactory performance.

This paper presents a new soft switching (SS) multiphase interleaved boost converter (MIBC) consisting of three boost converters (BCs) that operate rapidly and efficiently to charge a lithium-ion battery for photovoltaic (PV) power system applications. This suggested converter can operate both active power switches at zero current switching (ZCS), and when switched off, it works at zero voltage switching (ZVS), which reduces switching losses and increases switching efficiency (SF). An interleaving structure is employed to decrease the conduction losses and reduce the input current ripple (ICR) and output voltage ripple (OVR) to increase the power rating. The suggested converter regulates the input and output power by modulating the phase shift in the pulse width modulation (PWM) technique using the maximum power point tracking (MPPT) algorithm via two closed feedback loops, one for a slower external loop and the other for a faster PWM control internal loop, similar to current mode control. The suggested converter has high voltage (HV) gain due to the coupled inductors (CIs), which act as a transformer. This magnetic design structure in both the steady-state and transient states also decreases the size and enhances the converter’s performance. Finally, the suggested converter was experimentally verified on a 285W prototype, and the maximum efficiency was found to be 98%, based on modeling and experimental data.

Skin cancer ranks as one of the most widespread and lethal types of cancer. Without early identification and intervention, there is a propensity for the disease to disperse among the different body parts. This primarily occurs because of the abnormal proliferation of skin cells, which is often triggered by exposure to sunlight. Despite recent advancements in deep convolutional neural networks, there remains a difficulty in concentrating on the semantically meaningful aspects of a lesion. Our study introduces an innovative methodology to tackle this issue that couples deep learning techniques with a soft attention mechanism for feature aggregation, followed by classification layers to exploit the immense capability of the soft attention mechanism in amplifying the significance of crucial features while mitigating the impact of other irrelevant features within neural networks. Our proposed approach achieves a remarkable improvement of 3% in classification accuracy over the EfficientNetV2B3 model, 2.3% over InceptionV3, and 1.7% over InceptionResNetV2, compared to their original pre-trained counterparts on the benchmark ISIC Archive dataset. Our best model is EfficientNetV2B3, which achieved the highest accuracy of 95.6%.

Vehicular Edge-Cloud Computing (VECC) paradigm has emerged as a viable approach to overcome the inherent resource limitations of vehicles by offloading computationally demanding tasks to remote servers. Despite its potential, existing offloading strategies often result in increased latency and sub-optimal performance due to the concentration of workloads on a limited number of connected Roadside Units (RSUs). Moreover, ensuring data security and optimizing energy usage within this framework remain significant challenges. To address these concerns, this paper proposes a comprehensive framework for VECC systems. A novel load-balancing algorithm is proposed to effectively redistribute vehicles among RSUs, considering factors such as RSUs load, computational capacity, and data rate. Additionally, a robust security mechanism is incorporated using the Advanced Encryption Standard (AES) in conjunction with Electrocardiogram (ECG) signals as encryption keys to enhance data protection during transmission. To further improve system efficiency, a novel caching strategy is introduced, enabling edge servers to store completed tasks, which in turn reduces both latency and energy consumption. An optimization model is also proposed to minimize energy expenditure while ensuring that latency constraints are satisfied during computation offloading. Given the complexity of this problem in large-scale vehicular networks, the study formulates an equivalent reinforcement learning model and employs a deep learning algorithm to derive optimal solutions. Simulation results conclusively demonstrate that the proposed model significantly outperforms existing benchmark techniques in terms of energy savings.

Accurately predicting the impact of hormonal therapy on Prostate Cancer (PC) lesions is paramount for effective treatment planning and monitoring. This study proposes a comprehensive framework, termed Multimodal Hormonal Therapy Response Prediction (MHTRP), that integrates Magnetic Resonance Imaging (MRI) imaging markers, pathology markers (i.e., Gleason score), clinical markers (Prostate-Specific Antigen (PSA)), and demographic markers (patient age). The integration of these diverse markers leverages existing literature and clinical recommendations to construct a robust, personalized prediction system for hormonal therapy outcomes in PC. To extract imaging markers, the proposed framework employs a series of preprocessing steps, including prostate delineation, lesion segmentation, super-resolution, and other techniques, to prepare the data for analysis by our Deep Learning (DL)-based Multibranch Multimodality MRI Feature Extractor (M3FE). The M3FE comprises multiple branches, each employing a dedicated DL model to extract features from a specific modality. Subsequently, a fully connected classifier is utilized to discern crucial features from the extracted characteristics. Incorporating clinical, pathology, and demographic markers alongside extracted imaging markers, we performed ensemble fusion using a weighted sum fusion algorithm to predict the hormonal therapy effect on PC. Testing was conducted on a cohort of 39 patients stratified by PSA levels. Our system demonstrated high efficacy, achieving an impressive overall accuracy, with a sensitivity of 97.5% and a specificity of 100%. Despite the small cohort size, these results are encouraging and indicate that the combination of radiomic features with other markers significantly enhances the prediction of hormone treatment outcomes in PC, potentially paving the way for more personalized management strategies.

This paper presents the development and application of a Digital Twin (DT) model for the optimization of micro-grid operations. With the increasing integration of renewable energy resources (RERs) into power grids, micro-grids are essential for enhancing grid resilience and sustainability. The proposed DT model, enhanced with Deep Reinforcement Learning (DRL), simulates and optimizes key micro-grid functions, such as battery scheduling and load balancing, to improve energy efficiency and reduce operational costs. The model incorporates real-time monitoring, service-oriented simulations, cloud-based deployments, “what-if” analyses, advanced data analytics, and security features to enable comprehensive management of DTs. An optimization scenario was conducted to evaluate the effectiveness of the DT and DRL in improving micro-grid performance. The results demonstrated significant revenue improvements: 81.7% for PPO and 56.12% for SAC compared to the baseline. These findings highlight both the promising potential of DT technology and the critical importance of incorporating DRL techniques into the DTs to improve system performance and resilience.

The accurate segmentation of stroke lesions is crucial for the diagnosis and treatment of stroke patients, as it provides spatial information about affected brain regions and the extent of damage. While conventional manual techniques are time-consuming and prone to errors, advanced deep learning models have shown promising results in medical image segmentation. Recently, several complex architectures, such as vision Transformers and attention-based convolutional neural networks (CNNs), have been introduced for this task. However, the question remains whether such high-level designs are necessary to achieve the best results for all segmentation cases. In this paper, we evaluated the performance of four types of deep models for stroke segmentation: 1) a pure Transformer-based architecture (DAE-Former), 2) two advanced CNN-based models (LKA and DLKA) with attention mechanisms, 3) a hybrid model that incorporates CNNs with Transformers (FCT), and 4) the well-known self-adaptive nnU-Net framework. We examined their performance on two publicly available datasets, ISLES 2022 and ATLAS v2.0, and found that the nnU-Net, with its relatively simple design, achieved the best results among all the models tested. Furthermore, we investigated the impact of an imbalanced distribution of the number of unconnected components in each slice, as a representation of common variability in stroke segmentation. Our findings reveal a potential robustness issue of Transformers to such variability, which may explain their unexpected weak performance. Additionally, the success of nnU-Net underscores the significant impact of pre- and post-processing techniques in enhancing segmentation results, rather than solely focusing on architectural designs. These findings suggest that proposed complex architectures may be task-specific and simpler models with appropriate pre-/post-processing pipeline can be equally or more effective in generalization across different tasks in medical image segmentation.

The increase of Credit Card (CC) fraud in recent years requires the development of fraud detection systems that are both efficient and robust. This paper explored the utilization of machine learning models, with a particular emphasis on ensemble methods, to advance the detection of CC fraud. We present an ensemble model that incorporates different classifiers to address the dataset imbalance issue that is present in most CC datasets. We employed synthetic over-sampling and under-sampling techniques in certain machine learning algorithms to tackle the same issue. To address the issue of under sampling, we implemented ensemble random sampling, which was subsequently followed by Two-Stage Thresholding (TST) to eliminate outliers. The well-known Synthetic Minority Over-sampling Technique (SMOTE) was used for over-sampling. Data was utilized to train the model and subsequently generate predictions by utilizing testing data following the pre-processing of the dataset. The proposed ensemble model outperformed conventional methods in addressing the challenges associated with CC fraud detection, as demonstrated by a comparative analysis. Our findings suggest that ensemble approaches are effective at fighting fraud, paving the way for more adaptive and resilient fraud detection systems.

With the increasing demand for digitization, Optical Character Recognition (OCR) systems play a vital role in digitizing physical manuscripts. Several methods have been successfully deployed in the OCR domain. However, they often face challenges when dealing with low-resource regional scripts because of the limited training data and complex structure of characters. In such a scenario, Siamese Network (SN) meta-learning offers a promising solution for this problem by enabling quick adaptation to new tasks with minimal training data. Despite the success of SNs in various classification tasks, the traditional SN architecture seeks a compelling upgrade to improve its ability to distinguish between similar-looking characters of regional scripts. In this research paper, we propose a novel Priority-Smart Network (PSN) framework for traditional SN architectures, which can easily be incorporated into existing CNN backbone and improve their ability to identify characters in low-resource regional scripts. Furthermore, we propose the Enhanced Differential Edge Detection (EDED) preprocessing strategy explicitly designed for OCR tasks. We rigorously investigate and evaluate three benchmark low-resource script datasets to establish the effectiveness of our proposed techniques. Our experimentation results showcase significant advancements in character recognition accuracy and robustness, emphasizing the potential of SN combined with the PSN framework and EDED strategy for improving OCR systems in low-resource script.

The rapid emergence and growth in the number of digital communication platforms have resulted in a previously unimaginable volume of text data suitable for sentiment analysis. Public sentiment extraction and interpretation definitive in various areas of marketing, politics, and customer service sectors. Existing approaches prove inadequate to address language complexity and subtlety. Consequently, more advanced analytical tools are required. Therefore, this study employs a BERT model in performing sentiment analysis on textual data associated with ChatGPT, an AI conversational tool OpenAI develops. The datasets utilized in this research are three, which are Data Collection, ChatGPT Sentiment Analysis Dataset, and ChatGPT App Reviews Dataset. They range from topics such as tweets to comprehensive reviews and thus provide diverse backgrounds for testing the model’s applicability. The BERT model is chosen for its strong processing abilities, especially due to its capacity to understand text nuances. It applies sentiment analysis in six training epochs, and the results indicated a model competent in accurate sentiment classification, irrespective of language and sentiment uniqueness. The overall performance was ousttanding, as evidenced by the model’s accuracy and its classification precision in both datasets. These findings have the potential to make the BERT model dominant in all sentiment analysis tasks. The analysis would prove instrumental in providing in-depth public opinion and sentiments made possible by the model’s credible results. In this stutdy, it is clear that BERT models can significantly improve machine learning towards better understanding and interacting in human language. These results also provide a basis for further research into these modalities across real-time applications or further modification that would develop the model to be exposed to different kinds of data.

This research presents a modified back reflector to increase isolation for a small ultra-wideband (UWB) multiple-input-multiple-output (MIMO) antenna. The rectangular patch and narrow line strip on the bottom side of the substrate, arranged diagonally, make up this modified back reflector. The characteristic mode analysis provides a physical comprehension of the operating events that occur during the antenna’s evolution. The suggested antenna maintains an isolation level of more than 20.7 dB and has an operating bandwidth from 2.74 GHz to 14 GHz with a small dimension of 28.5 mm $\times 28.5$ mm $\times 1.6$ mm. Important factors that assess the effectiveness of the antenna are looked into, including radiation patterns, peak gain of 4.2 dB, radiation efficiency >85%, and MIMO diversity parameters. The suggested MIMO antenna is a feasible choice for UWB applications, as evidenced by the strong agreement between the measured and simulated results.

Many advanced techniques efficiently harvest the global maximum power (GMP) of the photovoltaic (PV) system, including machine learning techniques and metaheuristic algorithms based maximum power point tracking (MPPT). Nevertheless, they have shortcomings such as sluggish convergence and local maxima power (LMP) trapping. Combining techniques improves productivity. This work proposes an intelligent nonlinear proportional-integral-derivative (NPID) controller coupled with hybrid salp particle swarm optimization algorithm (SPSOA) to successfully harvest the GMP of PV system. The SPSOA-NPID controller’s performance is measured in regard to settle time, rising time, overshooting, peak time, undershoot, rotor rotation speed, and GMP under varied realistic irradiation and temperature profiles. In this work, the optimum parameter settings for the proposed NPID and the basic PID controllers were obtained utilizing the hybrid genetic algorithm and PSO (GA-PSO) techniques. Simulation findings proved the success and robustness of the SPSOA-NPID-based MPPT controller followed by GA-PSO-PID, GA-PID, GA-NPID, PSO-PID, PSO-NPID, P&O and INC respectively. The proposed SPSOA-NPID method obtained an average efficiency of (0.9946), the lowest average ripples (8.214 W), and the fastest average tracking time (0.052 s). Lastly, a hardware-in-loop (HIL) experimental carried out to confirm that the proposed SPSOA-NPID control can be practically implemented. Consequently, it is determined that the proposed SPSOA-NPID based GA-PSO control system is a potential MPPT approach based on the thorough research that have been given.

The difficulty of detecting liver disease at an early stage goes back to its limited number of symptoms. In this study, single and ensemble machine learning (ML) algorithms are applied to the Indian Liver Patient Dataset (ILPD) dataset, and their results, without and with feature selection techniques, are compared between each other and to the existing studies. Also, a two-level ensemble stacking model is applied based on several meta-ensemble classifiers and the feature selection technique to optimize the accuracy of the ensemble classifiers. Several data preprocessing techniques are employed to optimize the accuracy of the proposed work, including data encoding, data cleaning, data scaling, data skewing transformation, data balancing, and feature selection. The choices of single model ML are logistic regression (LR), K-nearest neighbors (KNN), decision tree (DT), linear discriminant analysis (LDA), and multilayer perceptron (MLP). In contrast, the choices of ensemble ML models are extra tree classifier, random forest (RF), gradient boosting, AdaBoost, extreme gradient boosting (XGBoost), and ensemble stacking classifier. Among the ensemble models, the ensemble stacking model achieved the highest accuracies (93.88% and 94.12%) when trained without and with the feature selection technique using the 10-fold cross-validation. The two-level ensemble stacking model achieved the highest performance with the metrics values: accuracy (94.01%), Precision (94.44%), Recall (94.25%), F1-score (94.01%), and area under the ROC curve (94.25%) when trained with feature selection technique. These results indicate that our proposed technique achieved a high prediction model for liver disease.

Quantitative Structure-Property Relationship (QSPR) modeling is one of the novel ways of predicting the physicochemical properties of a drug through its molecular descriptor (topological index (TI)). This study aims to predict the physical properties of antibacterial drugs by utilizing neighborhood sum degree TI (NTI) and employing two regression models namely, simple linear regression (LR) and support vector regression (SVR) as a machine learning model. In contrast to traditional methods of computing TI, we utilize the simplified molecular-input line-entry system (SMILES) of antibacterial drug compounds to compute the numerical parameters linked to antibacterial drug compounds. To enhance the predictive power of the QSPR model, we employ backward elimination method for LR model while incorporating sequential backward search selection (SBSS) technique for the SVR model. The study demonstrated that the SVR model outperformed the LR model, showcasing the SVR model’s ability to handle small datasets effectively. Hyperparameter tuning and the SBSS method were crucial in enhancing the SVR model’s performance by iteratively excluding non-informative features, thereby reducing error metrics and optimizing the model. The study identified the predictive capability of NTIs for various properties, including boiling point, melting point, flash point, enthalpy of vaporization, molar refraction, polarization, surface tension, and molar volume. The optimized SVR model showed high prediction accuracy for drugs such as Tobramycin, Gentamicin, Linezolid, Moxifloxacin, and Cilastatin, with predicted values closely matching actual physical properties. The results of our research highlight the significance of integrating advanced machine learning algorithms into drug property prediction leading to the optimization of drug design and discovery process.

LoRaWAN (Long Range Wide Area Network) technology has emerged as a leading solution for wireless communication in the Internet of Things (IoT) domain. With forecasts suggesting an unprecedented 500 billion IoT devices interconnected by 2030, the demand for efficient IoT applications has never been greater. LoRa technology offers low-power, long-range data transmission, addressing key requirements of IoT applications such as smart cities, agriculture, and industrial automation. This paper explores the challenges and advancements in LoRaWAN networks, focusing on Quality of Service (QoS) requirements. While traditional research efforts have primarily addressed network performance and energy optimization, the heterogeneous nature of IoT applications presents new challenges. The absence of QoS guarantees in legacy LoRaWAN protocols necessitates novel approaches to address diverse application needs. Moreover, the reliance of LoRaWAN on the Aloha access scheme introduces fundamental constraints for QoS consideration, making it difficult to develop complex algorithms for traffic control. Through a comprehensive review of existing algorithms, this paper evaluates strategies for managing QoS in heterogeneous LoRaWAN networks. By synthesizing the strengths of multiple algorithms, our proposed approaches, MixRA-opt and MixRA-H, aim to optimize network capacity and ensure satisfactory service levels for all applications, especially for large-scale, heavy-load scenarios. MixRA-opt is an optimization approach, while MixRA-H is a heuristic approach that serves as a simple approximation of the optimized method. The study includes simulations to validate the effectiveness of the proposed algorithms and provides insights for future research directions.

Enhancing model performance in agricultural image analysis faces challenges due to limited datasets, biological variability, and uncontrolled environments. Deep learning models require large, realistic datasets, which are often difficult to obtain. Data augmentation, especially through Generative Adversarial Networks (GANs), has become essential in farming applications, generating synthetic images to improve model training and reduce the need for extensive image collection. This review explores various GAN approaches for image augmentation in farming, investigating their challenges and limitations. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, 128 publications were analyzed to identify research trends and gaps in GAN applications within the farming industry. Key applications of GANs include plant classification, weed detection, animal detection and behavior recognition, animal health and disease analysis, plant disease detection, phenotyping, and fruit quality assessment. Persistent issues like limited training datasets, occlusion challenges, and imbalanced data hinder model performance across these applications. Recognizing these challenges is critical for enhancing the efficiency and effectiveness of farming operations. Finally, this review concludes with insights and future directions to foster progress in this field.

Smart grids (SGs) are technology-powered electricity networks that support bidirectional power and data flows. This allows real-time monitoring of demand and enhances the grid’s capability to dynamically adjust the generation and reduce the gap between supply and demand. However, implementing a smart grid in the power network comes with its own set of security challenges, such as cyber-security, distrust in participants, and lack of customer engagement due to various cyber-attacks. Such cyber-attacks will create distrust among consumers/prosumers to adopt the smart grid and distributed energy resources (DER) framework. To circumvent this, a blockchain-supported hybrid authentication and handshake algorithm (BSHAHA) for smart grids is proposed in this work, which authenticates data communication between peer-to-peer, aggregators, virtual power plants, and the grid. The algorithm was developed incorporating elliptic-curve cryptography (ECC) and advanced encryption standards (AES) to enhance privacy and session security. The proposed algorithms are verified and tested using formal cyber-security tools, such as the Random oracle model and AVISPA as well as by informal security analysis. Furthermore, to simulate a real-time test environment, this paper utilized ns-3 network simulator to simulate different smart meter scenarios, and the proposed algorithms are tested for power consumption and scalability, and results are presented. Moreover, blockchain simulation was first done in the local blockchain using Ganache and Truffle IDE and later using the Holesky Ethereum test network and remix-IDE. Lastly, this paper presented a comparison analysis of power consumption for different consensus mechanisms.

The Centralized Radio Access Networks (CRAN) decentralizes data and control planes by separating the baseband unit (BBU) from the central office, enabling energy-efficient “green networks” through the shutdown of underutilized BBUs. Analyzing extensive Call Detail Records (CDR) as big data, collected by service providers, has gained traction for extracting network features and studying activities. Thus, big data analytics are deemed as potential techniques that various research proposed to analyze the CDR. This paper introduces an energy-efficient CRAN network architecture based on the CRAN framework, focused on an innovative remote radio head (RRH)-BBU assignment. The objective is twofold: minimizing power consumption by deactivating underutilized BBUs and reducing inter-BBU handover rates based on CDR insights. In literature, the problem of assigning RRH to BBU is described as hard nonlinear programming (NLP) problem (bin packing, mixed integer), different suboptimal algorithms have been proposed to offer suboptimal assignment. This study employs clustering techniques to divide the complex NLP problem into simpler optimization tasks, achieving optimal RRH-BBU assignments. The proposed algorithm’s effectiveness was assessed using Milan city CDR as a case study, and its performance was validated against Milan’s land use map. The results indicated a remarkable 28.8% reduction in power consumption, alongside improvements in inter-BBU handovers.

Reinforcement Learning (RL) seeks to develop systems capable of autonomous decision-making by learning through interaction with their environment. Central to this process are reward engineering and reward shaping, which are essential for enhancing the efficiency and effectiveness of RL algorithms. These techniques guide agents toward desired behaviors, improve learning stability, and accelerate convergence by addressing challenges such as sparse and delayed rewards. However, the complexity of real-world environments and the computational demands of RL algorithms remain significant obstacles to broader adoption. Recent advancements in deep learning have enabled RL to handle high-dimensional state and action spaces, facilitating applications in robotics, autonomous driving, and complex decision-making tasks. In response to these developments, this paper provides one of the first comprehensive reviews of reward design in RL, with a focus on the methodologies and techniques underpinning reward engineering and shaping. By introducing a detailed taxonomy, critically analyzing current approaches, and highlighting their limitations, this work fills an important gap in the literature, offering insights into how reward structures can be optimized to meet the growing demands of modern AI systems.

The advent of big data has brought about significant opportunities and valuable insights to modern organizations. However, the proliferation of governance challenges is increasing dramatically, which necessitates searching for innovative approaches for managing, securing, and complying with government and organizational policies. This study conducted a systematic literature review and explored the profound impact of big data governance on modern organizations and highlighted the existing prevalent challenges and evaluated the mitigation approaches. We obtained the results of our defined research questions from peer-reviewed articles, industry reports, and academic theses published between (2019-2024). The results of our findings are categorized into four fundamental components of big data governance (i.e., people, process, data, technology), through illustrative tables and figures. Furthermore, this study has identified the need for effective and adaptive frameworks and models that can respond to emerging technologies and big data governance practices. After identifying the existing approaches for harnessing and mitigating the existing big data governance challenges, this study has proposed a new hybrid approach that integrates the viable system model and big data governance process model. The proposed approaches not only mitigate the existing challenges, but also serves as a diagnostic tool for identifying the maturity level of data governance and pinpoint the areas for improvement. However, the practical implementation of the proposed approaches remains a limitation of this study, urging avenues for future studies to explore its applicability and impact on modern organizations.

The growing interest in wind power and its expanding integration into the electrical power grid necessitate significant efforts to enhance the efficiency and performance of wind power generation Systems (WPGS) through optimized design and operation. This paper investigates the implementation of five optimization algorithms:sine cosine algorithm (SCA), grey wolf optimizer (GWO), particle swarm optimizer (PSO), transient search optimization (TSO), and a hybrid sine cosine algorithm-transient search optimizer (HSCATSO), to optimally design control schemes (ODCSs) of frequency converters within grid-connected WPGS. These schemes are crucial for improving conversion efficiency and performance. The optimal design procedure involves all optimization algorithms to minimize the total control scheme error by identifying the best settings for six controllers in the proposed system. The feasibility of ODCSs is validated through extensive simulation analysis in the MATLAB/Simulink environment, considering the challenges posed by the low wind speed profiles of Malaysia and severe grid disturbance conditions. The findings from the comparative analysis show that the ODCS developed with HSCATSO achieved the highest conversion efficiency, reaching 98.57%, and exhibited superior system stability, followed by the ODCS utilizing SCA. In contrast, the ODCSs designed using PSO demonstrated the lowest conversion efficiency, at 93.79%, along with the least system stability. This outcome confirms the significant role of ODCSs in improving the efficiency and performance of WPGS under the mentioned challenges, while also revealing that their effectiveness varies depending on the precision of design established by the utilized optimization algorithm.

The energy produced from various sources in modern power systems needs to be optimally planned for scheduling and managing the power system under specific conditions. In recent times, the world’s growing population, increasing energy needs, technological advancements, environmental concerns, and global climate change have led to a rise in the demand for electric energy. One significant solution to meeting this energy demand is the use of renewable energy sources (RESs) in power systems. However, integrating RESs has made the construction of power systems more complex. Moreover, Plug-in electric vehicles (PEVs) stand out as a highly promising technology for reducing carbon emissions in the transportation sector, aligning with the global Net-zero target. The optimal power flow (OPF) is a challenging problem when dealing with integrating renewable energy sources (RESs) and plug-in electric vehicles (PEVs) into modern power systems. In this paper, we address the OPF problem that arises from power systems with a high penetration of controllable renewable sources and PEVs. We simulate uncertain solar irradiance, wind speed, and PEV behavior using Weibull, lognormal, and normal probability distribution functions (PDFs) respectively. The proposed stochastic OPF problem, which incorporates controllable renewable sources and Vehicle-to-Grid (V2G) technology, is addressed using the hybrid metaheuristic optimization algorithm (AHMRFO). The AHMRFO hybrid algorithm combines the exploratory power of the manta ray foraging optimization (MRFO) with the exploiting strength of artificial hummingbird algorithm (AHA) resulting in a more balanced, robust, and efficient optimization tool. It can effectively handle complex problems, avoid premature convergence, and provide high-quality solutions across a wide range of optimization challenges. Initial evaluations for the proposed algorithm are executed on 7 Congress on Evolutionary Computation (CEC) test functions and its obtained results was compared with recent optimization algorithms such as original MRFO, AHA, supply demand-based optimization (SDO), hunter prey optimization (HPO), weighted mean of vectors (INFO), and northern goshawk optimization (NGO). Moreover, simulations carried out on different test systems demonstrate the effectiveness and efficiency of the proposed AHMRFO algorithm when compared to other algorithms such as the MRFO and AHA. The AHMRFO technique proves to be highly effective, yielding the lowest fitness values of 781.5836 ${\$}$ /h and 808.3611 ${\$}$ /h in the first two scenarios for the modified IEEE 30-bus system with RES and it has the minimum fitness values of 790.2783 ${\$}$ /h and 819.6573 ${\$}$ /h in their respective scenarios for the modified IEEE 30-bus system with stochastic Vehicle-to-grid and RES. These results highlight the precision and strength of AHMRFO technique in effectively addressing various instances of the OPF problem. Additionally, the Wilcoxon signed-rank test has been employed to demonstrate the superiority, effectiveness, and robustness of the proposed AHMRFO algorithm.

Digit-serial arithmetic has emerged as a viable approach for designing hardware accelerators, reducing interconnections, area utilization, and power consumption. However, conventional methods suffer from performance and latency issues. To address these challenges, we propose an accelerator design using left-to-right (LR) arithmetic, which performs computations in a most-significant digit first (MSDF) manner, enabling digit-level pipelining. This leads to substantial performance improvements and reduced latency. The processing engine is designed for convolutional neural networks (CNNs), which includes low-latency LR multipliers and adders for digit-level parallelism. The proposed DSLR-CNN is implemented in Verilog and synthesized with Synopsys design compiler using GSCL 45nm technology, the DSLR-CNN accelerator was evaluated on AlexNet, VGG-16, and ResNet-18 networks. Results show significant improvements across key performance evaluation metrics, including response time, peak performance, power consumption, operational intensity, area efficiency, and energy efficiency. The peak performance measured in GOPS of the proposed design is $4.37\times $ to $569.11\times $ higher than contemporary designs, and it achieved $3.58\times $ to $44.75\times $ higher peak energy efficiency (TOPS/W), outperforming conventional bit-serial designs.

Load flow analysis is essential for the effective operation and planning of power systems. While non-linear methods provide high precision, their computational complexity poses challenges for real-time applications, making them less practical for dynamic environments. In contrast, linear methods, while computationally efficient, often sacrifice accuracy, which can lead to suboptimal decision-making in power system management. This paper introduces a novel hybrid linear approach that synergistically combines two established linear load flow methods to overcome these limitations. By integrating the strengths of both methods, the proposed technique significantly enhances the accuracy of load flow analysis without compromising computational efficiency. To validate the effectiveness of the hybrid approach, extensive testing was conducted across various distribution networks. The results demonstrate notable improvements in accuracy compared to traditional linear methods, making this hybrid approach a promising solution for load flow analysis in practical applications. This advancement not only contributes to the field of power systems but also paves the way for more reliable and efficient real-time operational strategies.

This study explores the role of sodium chloride (NaCl) in enhancing the efficiency of microbial inactivation using microwave energy with two distinct methods: temperature-controlled treatment and predefined cycle. The primary objective is to maintain a stable temperature within the biological samples during treatment. Detailed descriptions of the treatment system’s design and the function of each element are provided. Simulations are conducted to illustrate the heating behavior of liquids with and without NaCl, revealing significant differences in temperature profiles between the two media. The experimental design and the objective behind each iteration are discussed. The experimental design includes three different power levels to assess their impact on disinfection rates, with durations determined based on extensive preliminary experiments to identify the most effective combinations for microorganism elimination. The experiments aim to compare the efficiency of the temperature-controlled method and the predefined cycle in maintaining desired temperature ranges and enhancing microbial inactivation. A key aspect of the study is the significance of exposure time in temperature-controlled runs. It is observed that exposure time critically influences the inactivation performance, necessitating the design of predefined cycles to achieve more consistent and efficient microbial inactivation. Additionally, in this investigation, the system effectively disinfects three types of microorganisms: bacteria, viruses, and yeasts. The study also examines treatments across three different media to understand the influence of NaCl presence on microbial inactivation efficiency. These insights emphasize the critical importance of optimizing exposure time and incorporating NaCl in microwave-based disinfection processes to achieve more effective microbial inactivation.

This scientific article aims to provide a comprehensive review of fault monitoring, diagnosis, and prognosis methods based on Artificial Intelligence (AI) for Hydroelectric Generator Units (HGUs). It presents a compilation of research studies that have utilized AI models for fault monitoring, diagnosis, and prognosis in HGUs. Additionally, it outlines the process for building an AI model in the context of fault management in HGUs and discusses the advantages and disadvantages associated with AI methods in this domain. Furthermore, the article examines the research prospects and trends of AI models for fault management in HGUs. By synthesizing existing literature and highlighting future directions, this article serves as a valuable resource for researchers and practitioners seeking to leverage AI techniques for effective fault management in HGUs.

Educational Data Mining (EDM) is the application of data mining methods in the educational domain. In the EDM field, we see mixed data (i.e., text and number data types). Grouping or clustering such data is challenging because determining the similarity between mixed data is poorly defined. Existing partition clustering algorithms for handling such data are based on two approaches: conversion of data types, where all data variables are converted to a single data type, and a mixed one, where the similarity measures of different data types are merged by either using a weighted sum approach as in Gower’s distance or by using mixed dissimilarity function as used in the k-Medoids algorithm to define a singular similarity measure for mixed data. Such a datatype conversion causes information loss, and this aspect is not discussed in the existing research literature. This study systematically reviews the past fifty-three years i.e. from 1971 to 2024 of research works on partition clustering algorithms applied to mixed data in EDM. A review of 104 research articles noted that most partitional clustering algorithms have continuous or categorical variables but not mixed variables. Researchers and practitioners often cite the lack of continuous and categorical variables analysis methods. Therefore, developing machine learning algorithms that can handle mixed data inherently present in the educational domain is increasingly becoming important. In addition to comparative analysis and analysis based on several factors, research gaps are also identified and mentioned in this article, and future insights are outlined.

Various computing approaches within vehicular networks, such as vehicular edge computing (VEC) and cloud computing, have been suggested to facilitate task offloading, aiming to improve user satisfaction. The features of vehicular networks, including the rapid movement of vehicles and the fluctuating distribution of vehicle densities, present challenges to task offloading with in the VEC. Numerous algorithms have been suggested to address these challenges and provide an effective task-offloading framework. This paper introduces a probabilistic model that analyzes task offloading across different computing tiers, alongside proposing a mobile computing paradigm tailored to the dynamic nature of vehicular networks (VN). This paradigm aims to maintain persistent connectivity and enhanced connection reliability despite mobility facilitating sustainable end-to-end service delivery. Building upon this premise, we propose a three-tier computing paradigm comprising Vehicle Edge of Things (VEoTC), VEC, and Cloud Computing (CC). Within the VEoTC tier, Service Vehicles (SV) equipped with computational resources serve as the mobile computing layer. The proposed model ensures continuous connectivity by extending the dwell time between the service requester and the vehicular computational resource. The model ensures that the relative speed between the service vehicle (representing computational resources) and the service requester remains constant while within the communication range. We proposed a probabilistic model for the end-to-end serving time of the proposed computing paradigm. Then, we computed the dwell time between the SV and the served vehicle based on real data published by Didi Chuxing GAIA Initiative for Chengdu city, China. Utilizing a simulated model, we illustrated the additional penalty incurred by the road side unit (RSU) handovers.

High frequency trading, often known as HFT, is a subset of algorithmic trading, which is one of the most significant improvements to the trading environment in recent years. Algorithmic trading gives traders the ability to trade or receive orders within extremely brief time intervals, such as minutes or seconds based. The very nature of high-frequency trading necessitates the use of high-speed data feeds, which deliver information in real time as well as trade executions for high-frequency traders. Due to the unpredictable nature of the market, several different methods are utilised in order to forecast the HFT time series. However none of these methods have been demonstrated to be a consistently acceptable forecast tool. The majority of HFT trading is conducted using traditional arbitrage-based methods.The conventional method of forecasting using time series does not have the capability to take into account seasonality or outliers in a dataset. The HFT time series data is consistently expanding from day to day, thereby becoming big and enormous. To solve this problem, we require a model that is both more precise and less time-consuming. The purpose of this paper work is to investigate historical models, contemporary models and hypothetical future models that can be used to anticipate HFT time series. Within the domain of HFT forecasting where accuracy and productivity are of the utmost importance, it also becomes critical to investigate state-of-the-art technologies. In light of the inherent difficulties that classical computing methods encounter when confronted with the intricacies of time series data, our focus shifts to quantum computing as a prospective resolution. This undertaking is driven by the imperative to confront prevailing obstacles in the field of forecasting, including enhancing precision, managing extensive datasets and alleviating the consequences of noise and uncertainty.The progression of quantum technologies is expected to bring significant changes that will enhance the reliability, precision and scalability of forecasting methodologies. The synergy between quantum computing and HFT forecasting holds the promise of reshaping how we approach and derive insights from time-varying datasets. The exploration and use of quantum computing for HFT forecasting marks a significant stride toward overcoming current limitations.

Today, vehicular communication can significantly improve traffic safety and road efficiency and as well as reduce the environmental pollution. The internet of vehicles (IoV) provides a communication platform between vehicles in intercity roads and streets of a city. Various messages such as safety messages and road traffic services can be transferred among objects of the platform including vehicles, Road Side Units (RSU), pedestrians, cloud’s servers and Trusted Authority (TA). To trust the messages and their initiators in IoV, the authentication mechanism is inevitable. On the other hand, vehicles insist on privacy preserving due to their security and safety. Therefore, an authentication mechanism between agents of IoV, which guarantees privacy preserving, is a hot topic in the research domain of IoV. This paper proposes a novel six-step authentication protocol, which is called PPA6-IoV, between vehicles such that the privacy of both initiator vehicle and receipt vehicle is preserved. During the authentication protocol, the session key for secure communication between two vehicles is established. The Crossware ARM simulator along with NS2 simulator is used to evaluate and compare the performance of PPA6-IoV with previous state-of-the-art techniques. Simulation results show that PPA6-IoV on average has 39% and 91% less communication and computational costs, respectively compared to previous state-of-the-art techniques. The results demonstrate that PPA6-IoV provides a proper level of both security and efficiency.

Fault detection systems support the operator, providing insight during the decision-making while having an (unknown) fault. Data-based models are a common option for a detection system. However, systems that rely purely on data-based models are normally trained with a specific set of data, which cannot necessarily prevent data drift. Thus, an anomaly or unknown condition detection mechanism is required to handle data with new fault cases. Besides, the model’s capability to adapt to the unknown condition is equally important to anomaly detection—in other words, its capability to update itself automatically. Alternatively, expert-centered models are powered by the knowledge of operators, which provides the models with production context and expert domain knowledge. The challenge lies in combining both systems and which framework can be used to achieve this fusion. We propose a novel adaptive information fusion methodology to define fault detection systems using evidence theory and uncertainty quantification. The main contribution of this paper is providing a general framework for the fusion of $n$ number of information sources using the evidence theory. The fusion provides a more robust prediction and an associated uncertainty that can be used to assess the prediction likeliness. Moreover, we provide a methodology for the information fusion of two primary sources: an ensemble classifier based on machine data and an expert-centered model. We demonstrate the information fusion approach using data from an industrial setup, which rounds up the application part of this research. Furthermore, we address the problem of data drift by proposing a methodology to update the data-based models using an evidence theory approach. We validate the approach using the Benchmark Tennessee Eastman while doing an ablation study of the model update parameters.

In the aviation industry, safety remains vital, often compromised by pilot errors attributed to factors such as workload, fatigue, stress, and emotional disturbances. To address these challenges, recent research has increasingly leveraged psychophysiological data and machine learning techniques, offering the potential to enhance safety by understanding pilot behavior. This systematic literature review rigorously follows a widely accepted methodology, scrutinizing 80 peer-reviewed studies out of 3352 studies from five key electronic databases. The paper focuses on behavioral aspects, data types, preprocessing techniques, machine learning models, and performance metrics used in existing studies. It reveals that the majority of research disproportionately concentrates on workload and fatigue, leaving behavioral aspects like emotional responses and attention dynamics less explored. Machine learning models such as tree-based and support vector machines are most commonly employed, but the utilization of advanced techniques like deep learning remains limited. Traditional preprocessing techniques dominate the landscape, urging the need for advanced methods. Data imbalance and its impact on model performance is identified as a critical, under-researched area. The review uncovers significant methodological gaps, including the unexplored influence of preprocessing on model efficacy, lack of diversification in data collection environments, and limited focus on model explainability. The paper concludes by advocating for targeted future research to address these gaps, thereby promoting both methodological innovation and a more comprehensive understanding of pilot behavior.

Fifth-generation (5G) networks consist of relatively smaller cells compared to legacy networks. Therefore, a user will take a shorter time toward the cell edge. This exposes the mobile station (MS) to frequent handovers, which are bottlenecks affecting the quality of service and user experience. In this study, we introduce a self-optimization method for three pivotal handover control parameters (HCPs): Threshold, Hysteresis and Time-To-Trigger. The proposed approach considers a holistic range of factors to determine the optimal values for these HCPs. These factors include the received power of reference signals (channel conditions), the speed and direction of the user (mobility profile), and the synchronization signal periodicity and handover procedure latency (representing system parameters). Through analytical deliberations, the study establishes a framework for achieving optimal HCPs to minimize the handovers, mitigate the ping-pong effects, reduce handover failures, and sustain a good throughput performance. Furthermore, considering the channel, user, and system parameters allowed cell-specific HCP optimization, enabling the implementation of this method with any of the measurement events outlined in the 3rd Generation Partnership Project (3GPP) release 16 for 5G. This study shows that concurrent self-optimization of Threshold, Hysteresis, and Time-To-Trigger can yield remarkable enhancement of handover performance.

Deep Learning (DL) plays a successful and influential role in medical imaging diagnosis, image detection, and image classification. Diabetes is a significant public health concern, and diabetic eye disease will be the leading cause of vision loss across the globe. This research proposed a multi-classification deep learning model for diagnosing and identifying four different diabetic eye diseases: Diabetic Retinopathy (DR), Diabetic Macular Edema (DME), glaucoma, and cataract which we called the DeepDiabetic framework. The proposed models were assessed using 1228 images from six different available datasets (DIARETDB0, DIARETDB1, Messidor, HEI-MED, Ocular, and Retina). In addition to the original dataset, we measured the performance of the deep learning models according to two different geometric augmentation methods called online augmented and offline augmented. The present work considers five architectures’ performances: EfficientNetB0, VGG16, ResNet152V2, ResNet152V2 + Gated Recurrent Unit (GRU), and ResNet152V2 + Bidirectional GRU (Bi-GRU). A comprehensive analysis and evaluation of these deep learning architectures is provided using public fundus datasets with four classes (i.e., DR, DME, Glaucoma, and Cataract). To the best of our knowledge, no other deep learning models for choosing between these models for these specific diseases are found in the literature. The results of the experiments showed that the EfficientNetB0 model outperforms the other four proposed models. The EfficientNetB0 model achieved 0.9876 in accuracy, 0.9876 recall, 0.9876 precision, and 0.9977 AUC based on fundus images. Our EfficientNetB0 model achieves 98.76% accuracy, while the previous studies only achieved 88.33%, 89.54%, 97.23%, and 80.33% accuracy, respectively. When compared to the previous studies as Fast-RCNN, RCNN-LSTM, and InceptionResNet, our EfficientNetB0 model achieves much higher accuracy, recall, precision, and AUC. According to the outcomes, our proposed models, especially the EfficientNetB0 model, are significantly more accurate than the state-of-the-art models.

Over the last twin decades, significant advancements have occurred in global electricity grids due to the widespread adoption of renewable energy resources (RES). While these sources play an essential role in total generation cost reduction, transmission power loss minimization, and reduction of environmental hazards related to traditional power plants. Still, however, the optimal planning and operation of the power system in the presence RES is considered a primary challenge due to the their stochastic natural and intermittency. One of the most complex and motivating issues in a power system is optimal power flow (OPF), a constrained optimization problem characterized by non-linearity and non-convexity. From these specifications, researchers competed in the past decades to find optimal solutions to stochastic OPF problems while keeping system stability. To tackle this challenge, an effective optimization algorithm which mimics on the foraging behavior of dwarf mongooses’ in the nature is introduced. The objective function considers reserve cost for overestimation and penalty cost for underestimation of intermittent renewable sources. To show the robustness and efficacy of the recommended optimizer, case studies on the customized IEEE 30-bus system and a realistic power system DZA 26-bus (isolated grid) are undertaken. Numerical findings show that the proposed DMOA beats all previous published-results and performs better over a variety of objective functions while finding high-quality optimally viable solutions. The obtained results demonstrate that the DMOA realized exceptional performance for both the test networks, with total generation cost minimized values of 780.982 ${\$}$ /h and 8283.942 US ${\$}$ /h, respectively. These results highlight the precision and robustness of DMOA in effectively addressing various instances of the OPF problem Furthermore, the one-way analysis of variance (ANOVA) test, a statistical approach, was employed to evaluate the superiority of the proposed algorithm and to highlight a certain level of confidence to our study.

The computing in the network (COIN) paradigm is a promising solution that leverages unused network resources to perform tasks to meet computation-demanding applications, such as the metaverse. In this vein, we consider the partial computation offloading problem in the metaverse for multiple subtasks in a COIN environment to minimize energy consumption and delay while dynamically adjusting the offloading policy based on the changing computational resource status. The problem is NP-hard, and we transform it into two subproblems: the task-splitting problem (TSP) on the user side and the task-offloading problem (TOP) on the COIN side. We model the TSP as an ordinal potential game and propose a decentralized algorithm to obtain its Nash equilibrium (NE). Then, we model the TOP as a Markov decision process and propose the double deep Q-network (DDQN) to solve for the optimal offloading policy. Unlike the conventional DDQN algorithm, where intelligent agents sample offloading decisions randomly within a certain probability, the COIN agent explores the NE of the TSP and the deep neural network. Finally, the simulation results reveal that the proposed model approach allows the COIN agent to update its policies and make more informed decisions, leading to improved performance over time compared to the traditional baseline.

Cervical cancer (CC), the most common cancer among women, is most commonly diagnosed through Pap smears, a crucial screening process that includes collecting cervical cells for examination. Artificial intelligence (AI)-powered computer-aided diagnoses (CAD) system becomes a promising tool for improving CC diagnosis. Deep learning (DL), a branch of AI, holds particular potential in CAD systems for early detection and accurate diagnosis. DL algorithm is trained to identify abnormalities and patterns in Pap smear images, such as dysplasia, cellular changes, and other markers of CC. So, this study presents a Computer Aided Cervical Cancer Diagnosis utilizing the Gazelle Optimizer Algorithm with Deep Learning (CACCD-GOADL) model on Pap smear images. The foremost objective of the CACCD-GOADL approach is to examine the image detection of CC. To accomplish this, the CACCD-GOADL methodology uses an improved MobileNetv3 model for extracting complex patterns in Pap smear images. In addition, the CACCD-GOADL technique designs a new GOA for the hyperparameter tuning of the improved MobileNetv3 system. For the classification and identification of cancer, the CACCD-GOADL technique uses a stacked extreme learning machine (SELM) methodology. The simulation validation of the CACCD-GOADL approach is verified on a benchmark dataset of Herlev. Experimental results highlighted that the CACCD-GOADL algorithm reaches superior outcomes over other methods.