Search Results (3,494)

Early fault detection in microsatellite thermal control systems (TCS) faces fundamental challenges: high-dimensional redundant telemetry channels, overlapping multi-scale periodicities that obscure anomaly signatures, and severely limited daily data downlink (1–2 passes per day) that restricts the temporal window for diagnosis. Existing data-driven approaches either rely on supervised learning, requiring labeled fault data that are scarce in practice, or employ univariate analysis that fails to capture inter-sensor spatial correlations. To address these limitations, this paper introduces a hybrid framework integrating information-theoretic feature selection and spatiotemporal deep learning. The Generalized Maximum Information Coefficient (GMIC) quantifies nonlinear dependencies between temperature channels for key channel selection, reducing dimensionality by 82% while preserving diagnostic information. A dual-level Seasonal Trend Decomposition (STL) method disentangles orbital-periodic dynamics from diurnal cycles, effectively isolating distinct thermal characteristics at multiple timescales. Each decomposed component is modeled using Convolutional Neural Network–Long Short-Term Memory (CNN-LSTM) networks to capture spatiotemporal dependencies for accurate temperature prediction. An adaptive threshold-based weighted error fusion mechanism enables early fault detection within a single day of telemetry data. Experimental validation on real satellite telemetry data demonstrates that the proposed framework achieves high-precision fault detection across multiple fault types using a minimal set of temperature channels, significantly outperforming existing benchmarks in both prediction accuracy and detection reliability. Full article

Objectives: Three stability-indicating analytical methods featuring outstanding sensitivity, selectivity, and precision were set up for the quantification of salicylamide (SAD) and ascorbic acid (ASC) in the presence of salicylic acid (SAL), which represents a possible impurity and degradation product of SAD. The aim was to develop sensitive, selective, precise, and eco-friendly assays appropriate for routine quality control of pharmaceuticals. Methods: Method (A) was a spectrophotometric technique of a successive derivative of ratio spectra built upon a two-step derivatization of ratio spectra utilizing double-distilled water as a solvent. SAD was quantified at 247.2 nm and 257.0 nm, and ASC at 251.8 and 259.8 nm, while SAL was quantified at 305.6 nm. Technique (B) relied on ratio spectra for the mean centering analytical process applied via two sequential stages, where the amplitudes derived after the second ratio spectra of the mean centering have been recorded on 291.0, 266.0, and 241.0 nm for SAD, ASC, and SAL, in that order. Method (C) involved TLC densitometric analysis, in which the separation was carried out upon plates of silica gel with chloroform–hexane–methanol–acetone–formic acid (5:3:2:1:0.2, in volumes) as a mobile phase, monitored by densitometric detection at 240 nm. The linear relationships were observed over concentration ranges of (0.2–2 µg/band) for SAD with ASC and (0.1–1 µg/band) for SAL. Validation of the presented techniques was performed in accordance with ICH strategies. Results: These developed techniques have been effectively analyzed for SAD with ASC in pharmaceutical dosage forms with non-interfering ingredients. A statistical comparison with the previously used HPLC technique revealed no considerable difference in terms of accuracy and precision. Greenness assessment using the AGREE platform produced scores of 0.72 for the spectrophotometric approach (benefiting from aqueous solvent) and 0.62 for HPTLC (limited by chloroform). Practical applicability (BAGI = 80 for both spectrophotometry and HPTLC) and overall quality indices (CACI = 83 for spectrophotometry; 80 for HPTLC) supported routine QC suitability. Conclusions: The three developed stability-indicating methods are accurate, precise, and selective for simultaneous assay of SAD and ASC in the presence of SAL and are suitable for quality control use. The spectrophotometric procedures combine high analytical performance with an improved environmental profile, while HPTLC offers comparable analytical reliability with slightly lower greenness due to organic solvent use. Full article

Software quality assessment in agile projects remains fragmented. Technical, process-related, and team-related indicators are often evaluated through separate models, tools, and reports. This fragmentation limits cross-project comparability and weakens evidence-based decisions for software quality improvement. To address this problem, this study proposes the Overall Software Quality Index (OSQI), a multidimensional decision-support framework for software quality assessment in agile projects. OSQI integrates code quality, process quality, and team quality into a single project-level assessment model. The framework was initially grounded in ISO/IEC 25010:2011 and is discussed in relation to the ISO/IEC 25010:2023 revision, particularly its explicit inclusion of Safety as a product quality characteristic. Since the industrial datasets used in this study were not collected from safety-critical systems, Safety was not modeled as a separate OSQI dimension in the current version; instead, it is addressed as a scope limitation and future extension. The measurement structure was defined using the Goal–Question–Metric (GQM) approach. An initial set of 49 candidate metrics was reduced to 15 core indicators. This reduction was performed using dimension-specific strategies: Random Forest-based feature importance for code quality, Delphi and Analytic Hierarchy Process (AHP) for process quality, and thematic consolidation for team quality. The selected indicators were normalized and integrated through entropy-based weighting. This process generates an interpretable composite quality score. The main contribution of OSQI is not the isolated use of these methods, but their integration into a reproducible and tool-supported framework. The framework converts heterogeneous software engineering signals into a unified decision-support index. OSQI was evaluated using industrial agile project data. The data included static code analysis outputs, issue-tracking records, team assessment results, and product outcome indicators. In an exploratory validation across five industrial projects, OSQI showed a strong positive association with Net Promoter Score ($r=0.97, p=0.0076$) and a strong negative association with churn rate ($r=-0.97, p=0.0061$). A supporting software tool was also developed to automate data integration, score calculation, visualization, and project-level comparison. The findings suggest that OSQI can support quality monitoring, project benchmarking, and evidence-based improvement decisions in agile software engineering contexts. Full article

Millimeter-wave radar enables non-contact monitoring of cardiac activity and therefore has the potential to reconstruct electrocardiogram signals without surface electrodes. However, existing radar-based electrocardiogram reconstruction methods still suffer from incomplete extraction of heartbeat-related information and insufficient modeling of electrocardiogram-related features, which limits reconstruction accuracy. To address these issues, this study proposes a millimeter-wave radar-based electrocardiogram reconstruction method that integrates a respiratory-harmonic-suppressed multi-channel signal-processing frontend with the proposed CA-WTBNet deep reconstruction network. First, based on maximal overlap discrete wavelet transform-based multi-resolution analysis, respiratory harmonics mixed into heartbeat-related components are suppressed by combining respiratory harmonic detection with a heart-rate frequency protection strategy, while cardiac-related information is preserved as much as possible. A multi-channel input representation is then constructed. Meanwhile, the proposed deep reconstruction network is developed to jointly model complementary channel-wise features, local waveform morphology, and temporal dependencies by integrating channel-attention mechanisms, convolutional residual modules, window-based Transformer blocks, and bidirectional long short-term memory. Experiments conducted on the public dataset show that our method achieves an average Pearson correlation coefficient of 0.9641, a mean normalized root mean square error of 0.0458, an average R-peak F1 score of 0.9956, and an average R-peak timing error of 3.13 ms on the test set. In comparison with related studies on the same public Resting dataset, the proposed method achieves the best overall performance among the compared methods, with a 0.53% improvement in Pearson correlation coefficient and a 10.20% reduction in normalized root mean square error over the best-performing compared method. Full article

Traffic accidents significantly impact societies and economies. The risk of collision is highest in urban areas, leading to devastating loss of life and escalating socioeconomic costs. In this context, numerous studies have focused on accurately predicting accident risk, severity, and duration using various methodologies. This paper presents an overview of traditional statistical models for accident prediction and a comprehensive systematic review of the literature on statistical modeling, machine learning (ML), and deep learning (DL) techniques employed in this field. Different methodologies and techniques are compared by categorizing studies that adopt similar approaches and analyzing them comparatively. Furthermore, a distinction is made between temporal and spatiotemporal models to describe how these approaches influence the accuracy of future predictions regarding accident occurrence and the duration of impact. This review distinguishes itself from similar works by not only comparing models and approaches, but also by analyzing how external features, such as meteorological data, road geometric design, and land usage, affect the probability of accidents and the models’ accuracy in forecasting road safety. The study explores the performance levels and limitations associated with a set of forecasting approaches, offering an analytical discussion of their differences and similarities, and potential future developments in this research space, including the use of hybrid models and reinforcement learning (RL). The results of this review indicate that DL models tend to be better suited to complex forecasting problems due to their superior ability to represent features and extract non-linear spatiotemporal correlations. This article concludes by describing various directions for further research, ranging from optimizing model architectures to integrating real-time big data into proactive prediction systems. Full article

In resource-constrained UAV swarm systems, multimodal sensory data are often affected by complex environmental factors, resulting in modality missing, signal degradation, and asynchrony, which significantly reduce the reliability of multimodal learning and incremental model updates. To address this issue, we propose a Compensatory Collaboration Modality-Balanced Sample Selection framework (CC-MBS), which improves robustness through modality quality modeling and cross-UAV collaborative compensation. Specifically, a modality confidence vector is introduced to quantify modality reliability from missing rate, degradation, and asynchrony. A lightweight collaboration mechanism is designed to exchange low-dimensional confidence information instead of high-dimensional features or model parameters. Based on the compensated confidence, a modality-aware sample selection strategy is further developed to prioritize high-value samples under limited memory. Experimental results in simulated UAV-swarm-inspired benchmark settings show that CC-MBS outperforms representation-based methods such as ShaSpec and its parameter aggregation variants (AVG, PFM, POW) in both modality compensation accuracy and communication–computation efficiency under missing conditions. In addition, it achieves stronger robustness than MBS and training-dynamics-based methods such as EL2N and GraNd in sample selection. These results demonstrate that CC-MBS effectively improves robustness and data efficiency for multimodal incremental learning under incomplete modalities. Full article

Barren plateaus (BPs) pose a major trainability challenge for quantum neural networks (QNNs) by causing gradients to concentrate near zero as circuit size, depth, or expressibility increases. This study presents a comparative benchmark of 10 BP mitigation strategies across six qubit settings (2, 4, 8, 12, 16, and 20) and three datasets of increasing complexity: Iris, MNIST, and MedMNIST. The evaluated methods include eight initialization-based strategies (Beta, Gaussian, Uniform Norm, CNN-based initialization, He-normal, He-uniform, Xavier-normal, and Xavier-uniform), one model-based variational encoder, and one optimization-based time-nonlocal Fourier parameterization. Experiments were implemented using PennyLane 3.10 and PyTorch 2.5 with simulator backends. We evaluate trainability using gradient variance and training loss, and we clarify that the benchmark analyzes simulated QNN optimization behavior rather than hardware-noise-resilient or noisy-label learning. Across the tested two-layer circuit configurations, the mitigation strategies maintained measurable gradient variance and stable loss reduction, suggesting that severe barren plateau behavior was not observed under the benchmark conditions. CNN-based and Beta initialization showed strong empirical behavior in variance retention and convergence speed, while Gaussian initialization was comparatively weaker in higher-dimensional settings. The study provides a reproducible benchmark structure for comparing BP mitigation behavior and identifies important limitations related to circuit depth, hardware noise, feature encoding, and classification performance that should be addressed in future QNN benchmarking. Full article

Peripheral blood metabolite concentrations vary with food intake and time of day, risking confounding effects in metabolomics studies with non-standardised sampling conditions or incomplete metadata. Such effects are often overlooked during study design, limiting the clinical translation of biomarkers and wasting resources for researchers, funders and clinicians. In our random sample of 100 human metabolomics studies, 56% did not control for food intake, and 59% did not explicitly control for sampling time. To provide a study design resource, we analysed a liquid-chromatography–mass-spectrometry-targeted dataset from controlled laboratory studies of 24 young, healthy participants (12 male, 12 female) sampled every 2 h for 34 h, with fixed-macronutrient meals provided at set times. Acute postprandial responses were quantified by effect size using pre- and post-meal windows, while daily rhythmicity was assessed using a mixed-effects cosinor model. Analyses were sex-stratified, and metabolites were classified as meal-responsive, time-of-day-responsive, both, or neither. Amino acids and their derivatives showed strong postprandial increases, whereas lipid classes showed minimal changes. Rhythmicity varied across metabolites, enabling the identification of features sensitive to meal timing and/or time of day. These results aim to provide a comprehensive dictionary of metabolite effect sizes for study design and metadata collection to support reproducibility and the clinical translation of potential biomarkers. Full article

Fatigue detection in remote tower air traffic controllers is important for maintaining operational safety under sustained visual monitoring and high cognitive workload. This study proposes MFD-Net, a dual-stream multimodal fusion framework using eye-tracking and electrocardiogram (ECG) signals. The model separately encodes eye-tracking and ECG-derived temporal inputs, incorporates an ECG-derived RMSSD expert feature, and performs lightweight late fusion for fatigue-state classification. Under the mixed-subject random-window protocol, MFD-Net achieved an Accuracy of 85.20%, a Recall of 83.33%, and an AUC of 0.9337. Because overlapping windows from the same participant and scenario could appear in both training and test sets, this result should be interpreted as a potentially optimistic within-distribution estimate. Under the stricter zero-shot leave-one-subject-out (LOSO) protocol, performance decreased substantially, with an Accuracy of $70.95\pm 21.59\%$, a Recall of $22.98\pm 36.30\%$, and an AUC of $0.6025\pm 0.2984$. This low zero-shot Recall indicates limited subject-independent fatigue-detection capability. Lightweight target-subject calibration and sequential probability aggregation improved adaptation and temporal stability, although the calibration results should be interpreted cautiously because random target-subject windows were used for fine-tuning. These findings suggest that eye-tracking and ECG fusion are promising under controlled conditions, while practical deployment requires deployment-oriented calibration protocols, recall-oriented optimization, and further real-world validation. Full article

Bio-based and biodegradable polymers such as short-chain-length (scl) poly(3-hydroxybutyrate) (PHB) and poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBV) are widely adopted in diverse areas such as healthcare, manufacturing, and packaging. However, high production costs and the complexity of tailoring their thermal properties, such as glass transition temperature (Tg), melting temperature (Tm), and crystallization temperature (Tc), hinder further adoption. The current study reported on the development of a raw dataset of PHB and PHBV materials compiled from 572 instances collected from the literature (558 instances) and in-house experiments (14 instances). The dataset encompassed compositional physicochemical parameters, molecular features, and corresponding thermal characteristics. After assessing data quality and filtering for completeness and available features, curated datasets were created for machine learning (ML) analysis. Two ML models, Random Forest (RF) and eXtreme Gradient Boosting (XGBoost), were utilized to predict values of Tg, Tc, and Tm using feature engineering methods that integrated chemistry-based descriptors with polymer-specific and experimental variables. The predictive performance of the models was systematically investigated using different combinations of input features to identify the most informative descriptor sets for each target property. The best-performing models were obtained using 118 data points for Tg and Tm and 201 data points for Tc, achieving R² values of 0.77, 0.76, and 0.82 for Tg, Tc, and Tm, respectively. Despite the reliable prediction of the thermal properties of scl-PHAs, the main limitations of the study were the relatively small dataset size for certain targets and incomplete or missing reporting of experimental conditions in the literature sources, which may introduce variability in the compiled data. The findings implied that curated polymer datasets and interpretable ML models can support the rational design of sustainable polymers with tailored properties for specific applications. Full article

Background: It is rare in clinical practice to encounter a common arterial trunk when the ventricular septum is intact. In this setting, other clinical diagnoses, such as hypoplastic left heart syndrome with aortic atresia, may be mistaken for a common arterial trunk. Data for this combination is largely limited to case reports and small case series. We have conducted a systematic review of reported cases, performing cluster analyses to provide an objective grouping of the cases. Methods: A systematic review of the literature was performed to identify cases of a common arterial trunk with an intact ventricular septum. Cases for which individual data were available were included in the final analyses. Cluster analysis using K-means clustering was conducted to provide an objective grouping of the hearts based on morphologic findings. Results: K-means clustering identified three distinct groups among hearts with a common arterial trunk with intact ventricular septum. The commitment of the common ventriculo-arterial junction to the left, right, or both ventricles was the defining feature of each group. Hearts with a common trunk committed to one of the ventricles demonstrated significant hypoplasia or atresia of structures related to the other ventricle. Conclusions: Distinct patterns can be identified when a common arterial trunk is found with an intact ventricular septum. They depend on the ventricle or ventricles, which support the common ventriculo-arterial junction. Full article

High-dimensional covariance estimation remains a fundamental challenge when the number of variables (p) substantially exceeds the sample size (n). In such settings, the sample covariance matrix is unstable, singular, and heavily contaminated by estimation noise. Although shrinkage estimators improve stability and thresholding methods promote sparsity, each approach alone may introduce bias or lose structural information. This study proposes a Noise-Adjusted Shrinkage Covariance (NASC) framework as a post-processing enhancement strategy for shrinkage-based covariance estimators. The framework first stabilizes the covariance structure through shrinkage toward a structured target, then suppresses noise-induced small covariance entries via thresholding, and finally applies a stabilization step to ensure positive definiteness of the resulting estimator. Sensitivity analyses were conducted to investigate the effects of the shrinkage and thresholding parameters, and the Monte Carlo simulations were subsequently performed using the best-performing parameter configuration. The simulation results showed that shrinkage alone may not sufficiently suppress entrywise noise, whereas NASC-adjusted estimators improved upon their corresponding shrinkage baselines in many scenarios, with the strongest gains observed for sparse covariance structures and for shrinkage estimators that do not explicitly suppress entrywise estimation noise. Improvements were more limited for highly optimized shrinkage estimators. Real-data analyses were conducted on the SRBCT and colon cancer benchmark datasets. On the SRBCT dataset, numerical stability and positive-definiteness properties were examined, while LOOCV-LDA classification performance without prior feature selection or dimensionality reduction was evaluated on the colon cancer dataset. The results suggest that NASC provides a computationally simple and numerically stable extension to classical shrinkage covariance estimation methods for high-dimensions. Full article

Background/Objectives: The “small-n-large-p” dilemma in mammography artificial intelligence (AI)—where the number of candidate imaging features far exceeds the number of labeled cases—commonly results in model overfitting, unstable feature selection, and poor generalization across clinical settings. This study aims to reassess the internal performance of the HiTopology-GOA-CSA (Grasshopper Optimization Algorithm–Crow Search Algorithm) feature-selection framework for mammography using a larger real Curated Breast Imaging Subset of Digital Database for Screening Mammography (CBIS-DDSM) cohort and a stricter leakage-aware evaluation strategy. Methods: In this retrospective computational study using public anonymized datasets, an expanded internal cohort of 98 CBIS-DDSM mass cases (49 benign, 49 malignant) was assembled from digital imaging and communications in medicine (DICOM) region of interest (ROI) series. A total of 1074 features were extracted per case, including 88 handcrafted radiomic descriptors and 986 EfficientNet-B5 deep features. HiTopology-GOA-CSA selected 102 features, corresponding to 91% feature reduction. Two internal evaluation modes were compared: Mode A, which matched the original pilot methodology by performing feature selection once on the full cohort before cross-validation, and Mode B, which used strict nested feature selection within training folds. Performance was assessed with 5-fold stratified cross-validation using a multilayer perceptron (MLP) classifier. Results: On the expanded cohort, Mode A achieved an area under the receiver operating characteristic curve (AUC) of 0.726 (95% CI: 0.594–0.858), sensitivity of 0.658, specificity of 0.651, and F1-score of 0.644. Under the stricter nested evaluation, Mode B achieved AUC of 0.683 (95% CI: 0.549–0.817), sensitivity of 0.598, specificity of 0.631, and F1-score of 0.595. Mean pairwise Jaccard similarity across nested folds was 0.604, indicating moderate feature stability. Benchmark comparisons showed that the proposed method was competitive but did not outperform standard baselines; LASSO logistic regression achieved the highest AUC of 0.739, while the proposed HiTopology-GOA-CSA + MLP achieved an AUC of 0.683. Real external validation on the locked VinDr-Mammo subset (n = 25) remained near-random (AUC of 0.500 [95% CI: 0.304–0.696]), with complete prediction collapse (sensitivity of 1.000, specificity of 0.000). Conclusions: The framework demonstrated feasibility for structured feature selection and stress testing in a small-cohort mammography AI setting; however, external validation revealed near-random discrimination and prediction collapse, indicating limited generalizability. These findings emphasize the need for benchmark comparisons, transparent uncertainty reporting, patient-level validation, and larger multicenter datasets before clinical translation. Full article

Background/Objectives: Survival prediction in non-small cell lung cancer (NSCLC) remains challenging, particularly in limited-sample settings where end-to-end deep learning models may suffer from limited generalization. This study aimed to develop a data-efficient, multimodal, and explainable framework integrating computed tomography (CT)-derived imaging information with clinical variables for NSCLC survival prediction. Methods: CT images, tumor segmentations, and clinical data from the publicly available NSCLC Radiomics (LUNG1) dataset (377 patients) were used. Tumor-focused regions were extracted using segmentation masks, and pretrained RadImageNet-InceptionV3 embeddings were obtained from the largest tumor-containing slice and neighboring-slice summaries. Deep imaging embeddings, engineered imaging features, and clinical variables were fused into a unified tabular representation. To improve robustness under limited-sample conditions, feature blocks were compressed using principal component analysis. CatBoost, XGBoost, and LightGBM models were trained on a development set and evaluated on a strictly held-out final validation set. Results: In three-class survival stratification, assigning censored/non-event patients to the upper survival group produced the strongest ordinal prognostic performance. Under the EX_PLUS_NON_EX_TOP setting, CatBoost achieved the best holdout score-based class C-index of 0.655. In continuous survival regression, LightGBM achieved the best holdout event-patient C-index of 0.576. Clinical variables provided the dominant prognostic signal, while compact deep image embeddings contributed complementary information, particularly in separating short- and long-survival groups. SHAP analysis confirmed contributions from both clinical and image-derived features. Conclusions: The proposed framework provides a proof-of-concept demonstration of a data-efficient and explainable image-to-tabular approach for NSCLC survival prediction under strict internal holdout validation. The results suggest that pretrained CT embeddings, clinical variables, gradient-boosted trees, and SHAP-based interpretation can be combined in a feasible, limited-sample survival modeling pipeline, while external validation remains necessary before clinical translation. Full article

Malware classification increasingly relies on machine learning models that combine static and dynamic evidence, yet their practical use is often limited by dataset inconsistency, high-dimensional feature spaces, and insufficient transparency. This paper presents an explainable hybrid malware-classification pipeline built on an aligned public dataset in which static and dynamic features are matched at sample level and share the same class space. The framework combines a Random Forest static branch, a calibrated XGBoost dynamic branch, and a weighted late-fusion stage whose branch weights are derived from inner-validation weighted-F1 rather than from test performance. On the corrected no-leak benchmark, static reduction compresses the static space from 771 to 258 features, while sparse-aggressive reduction compresses the dynamic space from 21,918 to 374 features. An early-fusion XGBoost baseline achieves the best multiclass aggregate scores, whereas the validation-weighted calibrated hybrid provides the strongest false-negative-first Benign vs. Malware profile, reaching malware recall 0.9998, benign recall 0.8053, and one false negative on the test set. The study shows that, once leakage is removed and fusion is validation-driven, the preferred hybrid architecture depends on the operational objective rather than on a single aggregate metric. Full article