Missing data are common in clinical prediction studies, yet their handling may affect not only predictive performance but also which model is judged best. This study examined the stability of classifier selection when missing data handling was changed under controlled, paired evaluation conditions. Using SUPPORT2 data from 9,105 patients, Logistic Regression, Random Forest, and XGBoost were evaluated with median, K nearest neighbor, and iterative imputation. The analysis retained natural missingness, added nested missing completely at random (MCAR) perturbations of 5%, 10%, 20%, and 30%, and included a separate 20% missing at random (MAR) condition. Repeated stratified five fold cross validation used the same patient partitions, model seeds, and artificial missingness masks across corresponding comparisons. Ranking stability was assessed through condition level winner changes, paired rank reversals, and agreement across receiver operating characteristic area under the curve (ROC AUC), Average Precision, and Brier Score. The condition level winner remained stable under natural and mild additional missingness, but became dependent on imputation at higher missingness. Winner changes occurred in 6 of 12 imputation comparisons and 9 of 15 missingness comparisons, while pairwise rank reversals occurred in about 44% of matched repeat comparisons. All three metrics selected the same winner in 8 of 18 conditions. Because competing winners were separated by small ROC AUC margins, the results indicate sensitivity of model selection rather than large performance advantages. Reporting ranking stability alongside conventional performance estimates may therefore provide a more cautious basis for comparative clinical prediction studies.

The rapid growth of Internet of Things (IoT) applications has increased the number of connected devices operating in shared wireless environments. IoT systems rely on sensors, actuators, embedded electronics, software, and internet connectivity to collect, exchange, and process data across sectors such as smart homes, healthcare, transportation, industrial systems, energy systems, and public infrastructure. However, the dense deployment of wireless IoT devices can create significant Electromagnetic Compatibility (EMC) challenges, especially when multiple transmitters and receivers operate in proximity or within the same or adjacent frequency bands. These conditions may lead to Electromagnetic Interference (EMI), which can degrade communication performance, reduce device reliability, and affect safety in critical environments. This study uses a regulatory review and conceptual framework development, drawing on literature synthesis, regulatory requirement extraction, analytical classification, regulatory mapping, and risk-based EMC categorisation. It identifies key EMC and EMI challenges in IoT applications, particularly in dense wireless and low-power communication environments. It analyses immunity requirements in critical sectors such as vehicular networks, healthcare, industry, and public infrastructure. It also develops a regulatory framework that defines stakeholder roles and classifies IoT connection methods, application areas, and device structures to support safer deployment and reduce interference risks.

Feature reduction is widely used in machine learning based phishing URL detection to lower model complexity while retaining detection performance. However, conclusions about how much reduction is acceptable may depend on how training and testing data are partitioned. This study examines whether conclusions about feature reduction remain stable when evaluation changes from random stratified to domain disjoint cross validation. Experiments were conducted on URL-Phish Version 2 using Mutual Information ranking within each training fold and four predefined feature budgets, from 22 to 5 features. Logistic Regression and Random Forest were evaluated under both protocols, with PR AUC as the primary metric. PR AUC decreased as the feature budget was reduced for both classifiers under both evaluation protocols, so the qualitative conclusion about feature reduction remained consistent within this experimental setting. However, the size of the protocol gap varied across classifiers and feature budgets: it was clearest for Random Forest under the smallest feature budget, whereas the Logistic Regression gaps were less clearly separated from fold-level variability. Domain disjoint evaluation also produced greater fold-level variability. The results show that claims about compact phishing URL representations should be interpreted together with the evaluation protocol used. Future work should examine whetherthis pattern persists across additional datasets, classifiers, and domain grouping rules.

The volume of Arabic text online keeps growing, and with it the need for systems that sort it into meaningful categories. A quieter problem often passes unnoticed: when long documents are split into shorter segments and pieces of the same document fall on both sides of the boundary between training and testing, the reported scores can look far better than the model deserves. This study treats that document level leakage as its central subject. Using a purpose built multidomain Arabic corpus of 1,000 segments drawn from 183 source documents across seven categories, we first measure how much leakage distorts the numbers. Under a naive random split, an optimized Linear SVM appears to reach 0.9800 accuracy; under a grouped split keyed to the document identifier, the same model reaches 0.8528. The gap of about 12.7 accuracy points is attributable to leakage, and across models the inflation ranges from roughly 11.9 to 30.4 points. Leakage also flattens the ranking, so a weak Naive Bayes looks almost as strong as the best model. Under the honest protocol, an ablation shows that character level TF IDF alone matches the combined word and character representation and slightly exceeds it on Macro F1, 0.7801 against 0.7713, while word features mainly refine precision. McNemar tests and a document clustered bootstrap show that the strongest linear models are statistically indistinguishable. A fastText baseline reaches 0.6497 accuracy, and two pretrained Arabic transformers, AraBERT and MARBERT, perform comparably to the linear model rather than surpassing it. A class level analysis exposes the General category as a structural confound whose removal raises accuracy to 0.9349. Overall, leakage aware evaluation paired with subword rich TF IDF and a Linear SVM forms a strong, honest, and practical baseline for Arabic text classification.

Task scheduling has often been treated as a secondary concern in fog computing, something to address only after the architecture is defined. This review argues that it is instead the central runtime decision in the fog to cloud continuum, because it determines whether the promised gains in latency, energy efficiency, and reliability can actually be achieved. The study synthesizes 102 foundational, methodological, and technical sources on task scheduling in fog enabled IoT environments. The aim was not simply to catalogue algorithms, but to examine how the field has framed the scheduling problem and how that framing has changed over time. The evidence reveals a clear progression. Early studies commonly assumed stable resources, predictable workloads, and simplified network conditions, which made scheduling easier to model but less representative of real deployments. More recent work has relaxed these assumptions and introduced dynamic, multi objective, application aware, learning based, and deployment oriented approaches. Six research streams emerge from this evolution. The main finding is that algorithmic sophistication has advanced faster than evaluation practice. Reported improvements in latency, energy consumption, and other QoS metrics are often difficult to compare because studies use different workloads, simulators, baselines, and experimental assumptions. Scheduling and orchestration overhead is rarely measured, while physical testbed validation remains limited. These gaps directly affect confidence in whether a proposed scheduler would behave as expected in operational fog systems. The review therefore identifies several priorities for future work: standardized benchmark workloads, cloud native scheduling that accounts for container lifecycle and microservice dependencies, resilience aware scheduling that treats failures and migration as first class concerns, and carbon aware orchestration that extends beyond energy minimization. Beyond the taxonomy, the paper argues for a shift from proof of concept scheduling studies toward reproducible, transparent, and deployable fog systems.

Pneumonia remains a major global health burden, where timely recognition on chest X ray images is clinically important yet often challenged by subtle radiographic signs and variability in interpretation. This paper presents a controlled comparative evaluation of four convolutional neural network architectures, MobileNet, ResNet50, VGG, and InceptionV3, for binary classification of chest X ray images into diseased and normal cases. Experiments were conducted using a publicly available Kaggle dataset of 4,479 images under a unified preprocessing and evaluation protocol. Performance was assessed on a held out test set of 300 images, including 200 diseased and 100 normal cases, using accuracy and macro averaged precision, recall, and F1 score, supported by confusion matrix analysis. The results show that MobileNet achieved the highest test accuracy at 95.0 percent, while ResNet50 and VGG achieved 94.7 percent, and InceptionV3 achieved 92.0 percent. Confusion matrix inspection indicates that MobileNet produced the fewest false negatives for diseased cases in this setting, which is important for screening oriented use. Inference time measurements using batch size 1 at 180 × 180 input on CPU further highlight the efficiency advantage of lightweight architectures. Overall, these findings provide a reproducible benchmark to support architecture selection for computer assisted pneumonia screening and clinical triage

Files