Search Results (654)

Parkinson’s disease (PD) is associated with alterations in gait symmetry and plantar loading that can be examined using vertical ground reaction force (VGRF) recordings. This study presents a confidence-stratified, physiology-anchored multiple-instance learning framework with concept-bottleneck-inspired pathways (implementation identifier: DRO-PAS-MIL-CBM; hereafter, PAS-MIL) for retrospective session-level PD-versus-control classification. Each gait session is represented as a bag of temporal windows. Eight predefined bilateral signal descriptors are combined with eight learned latent temporal dimensions, aggregated through attention-based pooling, and processed by concept-guided, prototype, anchor-only, and static-feature expert pathways. The evaluation used five-fold person-grouped cross-validation on 306 sessions from 165 participants in the PhysioNet Gait in Parkinson’s Disease database.Inner person-grouped out-of-fold ExtraTrees probabilities were used to construct the confidence strata and distillation targets. PAS-MIL achieved a pooled session-level area under the receiver operating characteristic curve of 0.771, average precision of 0.890, and a mean fold AUC of $0.826\pm 0.041$. Relevance analysis identified C05 (asymmetry variability) and C08 (bilateral change mismatch) as the highest-weighted predefined physiological anchor descriptors. Protocol-stratified sensitivity analysis showed variation across the three source sub-studies, with AUCs ranging from 0.740 to 0.790. Probability calibration remained suboptimal after temperature scaling (mean per-fold ECE, $0.291\pm 0.042$). The results demonstrate the feasibility of integrating physiology-informed descriptors, temporal representation learning, and session-level aggregation. The study is a retrospective proof of concept and does not establish external robustness or clinical deployment readiness. Full article

Vehicular traffic is a major source of air pollution; however, the contribution of remotely acquired traffic information to local machine-learning (ML) air-pollution models remains insufficiently characterised. This study evaluates four interpretable tree-based ML models (Random Forest, Extra Trees, LightGBM, and XGBoost) under six predictor scenarios combining progressively larger predictor sets, ranging from remotely acquired traffic, meteorological, and temporal variables alone to the inclusion of measurements from one and four neighbouring monitoring stations, to estimate NO${}_2$, PM${}_{10}$, PM${}_{2.5}$, and O${}_3$ concentrations across several sites in London. ML model performance was compared with a ridge linear regression model as a baseline, with spatial interpolation methods and with a cross-site validation experiment. When modelling without data from neighbouring stations, the RMSE for NO${}_2$ ranged from 9.73 to 11.66 $\mathsf{\mu }$g/m${}^3$ without traffic information, compared with 8.72 to 11.52 $\mathsf{\mu }$g/m${}^3$ when traffic information was included. Additionally, for NO${}_2$, SHAP analyses indicate that traffic-related variables can contribute at levels comparable to pollutant measurements from neighbouring monitoring stations in traffic-dominated environments. Full article

This study proposes an innovative framework that combines hydroclimatic and agro-environmental predictors, including nitrate concentration and NDVI, with climatic factors such as Rainfall, temperature, and evapotranspiration to forecast piezometric level variations in the Saïss Basin, Morocco. Eight Machine Learning (ML) models were benchmarked, and feature importance was assessed using Shapley Additive exPlanations (SHAP) to ensure model transparency and interpretability. In parallel, the PELT algorithm was applied to detect structural change points, while Sen’s slope estimator and the Mann–Kendall test were used to quantify long-term trends. The Extra Trees (ET) model achieved the best performance (R² = 0.92), with Rainfall emerging as the most influential predictor, followed by nitrate concentration, confirming the added value of hydrochemical indicators for groundwater forecasting. Change-point analysis revealed significant declines during the 1980s and 1990s, followed by lower-amplitude fluctuations since the late 2000s. Projections toward 2050 suggest partial stabilization in the central part of the basin under favorable recharge conditions, whereas persistent declines are expected to continue in peripheral areas subjected to sustained groundwater abstraction pressure. These findings provide a robust and transferable decision-support tool for the sustainable management of groundwater resources in semi-arid agricultural area. Full article

Certified seed purity is a prerequisite for sustaining the agronomic performance and commercial value of Khao Dawk Mali 105 (KDML-105), Thailand’s premium aromatic rice; however, conventional inspection methods are destructive, labour-intensive, and poorly suited to high-throughput operations. This study developed a non-destructive purity inspection system integrating five-band MSI acquired using a MicaSense RedEdge-MX sensor with three machine learning classifiers: Extra Trees (ET), Random Forest (RF), and Support Vector Machine (SVM). The system was evaluated systematically across four illumination levels (360.90–13,188.46 lx) to discriminate KDML-105 from three morphologically similar contaminating varieties: Chainat-1, RD-6, and RD-15. Two-way ANOVA confirmed that classifier type was the dominant performance determinant (η² = 0.847). Under optimal illumination (L4, 13,188.46 lx), ET achieved the highest accuracy (88.8%), recall (86.6%), and F1-score (88.4%) with a training time of 0.098 s. External validation confirmed model generalisability: KDML-105 seeds were identified with 90.6% accuracy and 95.6% recall, while Chainat-1 and RD-6 yielded accuracies of 85.1% and 90.6%, respectively; RD-15 remained challenging (62.9%; 67.4%) owing to spectral proximity to KDML-105. These findings establish MSI combined with ET classification as a viable, cost-effective approach for automated seed purity screening in certified rice production. Full article

Objective: This study aimed to develop machine learning (ML)-based models for the early prediction of macrosomia using only maternal sociodemographic and obstetric data. Methods: This retrospective study included 100 pregnant women who delivered at the Obstetric Clinic of Eskisehir City Hospital between January 2022 and December 2023. Participants were classified as nulliparous (n = 48) or parous (n = 52) and further categorized according to the presence or absence of fetal macrosomia (birth weight > 4000 g). Predictor variables included maternal age, body mass index (BMI), gravida, smoking status, history of diabetes mellitus, and hypertension; previous birth weight was additionally included in the parous model. Separate machine learning models were developed for nulliparous and parous women using Extra Trees Classifier, Light Gradient Boosting Machine (LGBM), eXtreme Gradient Boosting (XGB) Classifier, Random Forest, and Logistic Regression. The dataset was randomly divided into training (80%) and testing (20%) subsets. Internal validation was performed using 10-fold cross-validation within the training dataset to optimize model performance and reduce overfitting. Given the relatively small sample size, this study was designed as a pilot exploratory investigation. Results: Overall, 48 nulliparous and 52 parous mothers were included in the study. Among the nulliparous women, 22 (45.8%) had macrosomic newborns, whereas 26 (54.2%) had normal birthweight newborns. Among parous women, 28 (53.8%) had macrosomic newborns, while 24 (46.2%) had normal birthweight newborns. For nulliparous mothers, the XGB Classifier achieved the highest accuracy (80%) and AUC-ROC (82.2%), demonstrating robust predictive performance. For parous mothers, the XGB Classifier again outperformed other ML models, achieving an accuracy of 72.7% and an AUC-ROC of 83.3% in predicting macrosomia. Conclusions: This study highlights the feasibility of ML-based decision support systems in obstetrics, particularly in low-resource settings, to predict macrosomia using readily available maternal characteristics. As a pilot study, these findings should be interpreted cautiously and require validation in larger multicenter cohorts before clinical implementation. Full article

Airborne bacterial diseases driven by pathogen aerosols in enclosed greenhouses spread rapidly, challenging traditional early-warning methods. This study developed a two-step monitoring system for cucumber angular leaf spot using low-cost particulate matter (PM) sensors, qPCR, and machine learning. Evaluated across spatially independent greenhouse trials using 732 plot-days of data, PM sensors were utilized as dynamic physical proxies alongside microclimate data. These proxies continuously estimated the fluctuations of pathogen aerosols suspended in the greenhouse air. When estimated aerosol risks exceeded a pathogenic threshold, targeted air sampling and qPCR quantification were triggered. For pathogen monitoring, the Extra Trees (ET) surveillance model accurately predicted the accumulation of airborne pathogen aerosols (R² = 0.884). For disease forecasting, by integrating the quantified aerosol loads with environmental factors, the XGBoost prediction model forecasted the daily disease index change rate with high precision (R² = 0.874). SHapley Additive exPlanations (SHAP) analysis confirmed that the concentration of airborne pathogen aerosols and vapor pressure deficit were primary drivers of disease expansion. By combining continuous physical sensing of greenhouse air with risk-triggered biological quantification, this framework provides a feasible strategy to partly compensate for the lack of biological specificity of PM sensors and supports early-warning management of airborne bacterial diseases in protected agriculture. Full article

This study examines the feasibility of predicting net repair costs (Estimated Cost, PLN) for multi-family residential buildings from data extracted from technical inspection reports. Rather than merely comparing algorithmic performance, the analysis was designed as a diagnostic sequence aimed at identifying the sources of prediction error. Four machine learning algorithms (Extra Trees, Random Forest, XGBoost, and GBM) were first applied to direct regression of repair cost. We then tested whether the difficulty of estimating exact cost values stems from the high variability of the target variable and whether this limitation can be mitigated by a two-stage approach: assigning observations to one of three cost-risk bands (Low, Moderate, High) and subsequently estimating cost within the assigned band. The empirical cost distribution was strongly right-skewed (median: 4500 PLN; mean: 63,402 PLN; maximum: 3,680,524 PLN). The best direct regression model achieved an R² of 0.452, while the best fully deployable two-stage model, combining an XGBoost classifier with a Random Forest regressor, achieved an R² of 0.392. When it was assumed that the actual cost-risk bands were known, an R² value of 0.839 was obtained, indicating that the main source of error is not regression within the bands, but rather the initial stage of assigning the bands. These results demonstrate that reporting a single global R² for highly skewed, weakly identifiable cost data can be misleading, and that decomposing predictive performance into band-assignment and within-band regression components provides a more informative evaluation. This article also points to concrete directions for improving the underlying database, particularly through the inclusion of variables describing repair quantity, unit of measure, and detailed repair scope. Full article

Background: Environmental exposures are biologically plausible contributors to thyroid cancer outcomes, but population-based evidence is heterogeneous and often focuses on incidence rather than stage at diagnosis. Objective: This study aimed to evaluate whether ecological state-year environmental indicators add predictive information for advanced-stage thyroid cancer at diagnosis in a large SEER-based cohort (U.S. National Cancer Institute’s Surveillance, Epidemiology, and End Results Program). Methods: We conducted a retrospective registry-linked analysis of 368,726 SEER thyroid cancer cases with valid combined summary stage. Advanced stage was defined as regional or distant disease and occurred in 118,509 cases (32.14%). Environmental variables were linked as state-year indicators using five-year moving averages from the years preceding diagnosis. Regularized logistic regression, gradient boosting, and Extra Trees models evaluated incremental predictive performance. Results: Adding environmental variables to demographic, socioeconomic, histology, and laterality predictors produced small random-split improvements across model families. However, after SEER registry was added before environmental variables, the remaining environmental increment was minimal. Registry-group holdout validation did not support improved geographic generalizability from the environmental feature set. Conclusions: State-year environmental indicators carried limited predictive information for advanced thyroid cancer stage, and much of this information overlapped with registry and geographic structure. These findings clarify both the potential and limitations of ecological environmental linkage in SEER-based prediction studies. Full article

Compost maturity screening supports safe land application and organic waste recycling, but germination index (GI) assays are not always available for rapid process assessment. This study performed a secondary GI-based maturity classification reconstruction using a published Nature Food composting dataset. From this source, 184 observations from 24 manure-based composting batch trajectories across five feedstock domains were retained when GI and routine physicochemical variables were available. GI values were converted into three maturity stages and a binary mature/non-mature endpoint, while the GI itself was excluded from model inputs. Logistic regression, random forest, and extra trees models were evaluated under random split, batch-aware group split, and leave-one-feedstock-domain-out validation. Random and group splits showed stronger apparent performance than cross-feedstock validation, indicating sensitivity to feedstock-domain transfer. In binary classification, the area under the receiver operating characteristic curve (ROC-AUC) remained relatively high in several model–domain combinations, whereas mature-class F1 declined, revealing a discrimination decision gap under default thresholds. Training-domain threshold calibration partially improved mature-class detection without using the held-out feedstock domain for threshold selection. These results support feedstock-aware validation and calibrated decision thresholds for sustainable compost maturity screening. Full article

Circular RNAs (circRNAs) participate in gene regulation and disease progression, but experimental mapping of circRNA-disease associations (CDAs) remains costly and incomplete. We developed NARVGA, a hybrid prediction framework that combines non-negative matrix factorisation (NMF) with an adversarially regularised variational graph autoencoder (ARVGA). Functional, semantic and Gaussian interaction-profile kernel similarities are integrated; K-means-derived co-membership graphs reduce diffuse similarity connections; ARVGA learns nonlinear topological embeddings; and NMF captures complementary low-rank association patterns. An Extra Trees classifier then scores candidate circRNA-disease pairs. In the original transductive stratified five-fold benchmark on CircR2Disease, NARVGA achieved an area under the receiver operating characteristic curve (AUC) of 0.9868, an area under the precision–recall curve (AUPR) of 0.9887, 94.32% accuracy and a 94.44% F1-score. Ablation analysis identified cluster sparsification as the largest individual contributor, with additional gains from adversarial regularisation and low-rank augmentation. Without task-specific retuning, the model obtained AUCs of 0.9503 on LncRNADisease and 0.9700 on HMDDv4. Performance was stable across GCN depths and cluster settings, and 19 of the 20 highest-ranked hepatocellular carcinoma candidates had supporting published evidence. NARVGA provides a within-network prioritisation tool for RNA-disease association studies, although prospective validation, hard-negative assessment and independent-cohort testing remain necessary. Full article

Accurate HVAC load identification from low-frequency smart-meter data is important for commercial-building demand response and energy management, but remains difficult when high-frequency measurements, detailed physical models, and HVAC submeters are unavailable. This paper proposes a hybrid data-driven and knowledge-driven framework that integrates HVAC load disaggregation with day-ahead forecasting. Operating modes are first identified from normalized daily load shapes and calendar features using K-medoids clustering. For each mode, a non-HVAC baseline is constructed from mode-wise low-load observations, cyclic smoothing, and a label-free shape correction based on the representative operating profile and lower-tail load dispersion. HVAC load is then obtained as the physically constrained residual between whole-building load and the corrected baseline. Historical disaggregation estimates are subsequently used as pseudo-labels for a leakage-controlled Extra-Trees forecasting model that combines target-day weather and calendar information with admissible historical load features. Experiments on three 15 min NREL ComStock commercial-building datasets show that the proposed disaggregation method achieves R² values of 0.8308–0.9246 across building types, while the proposed forecasting model attains an R² of 0.7812 on the held-out test period. The results demonstrate an interpretable and submeter-free approach for HVAC load analysis under low-frequency metering, with the strongest performance under cooling-dominated conditions. Full article

Background: Acute myeloid leukemia (AML) is a heterogeneous and aggressive hematologic malignancy, and LY-11h is a novel acylhydrazide-based histone deacetylase inhibitor with promising therapeutic potential for AML. However, its poor aqueous solubility, limited intestinal dissolution, and narrow therapeutic window hinder oral formulation development and motivate the development of dosage forms with flexible dose-design capabilities. Herein, an integrated hot-melt extrusion (HME)–fused deposition modeling (FDM) strategy was developed to convert LY-11h into printable amorphous solid dispersion (ASD) dosage forms. Methods: HPMC-AS was used as a pH-responsive carrier to enhance intestinal release while restricting premature gastric release, and HPC-EF was incorporated to improve filament processability. Single-factor and DoE studies identified critical formulation and process variables and established formulation–process–property relationships, while machine learning further modeled nonlinear interactions and guided optimization. In-line near-infrared spectroscopy combined with polarized light microscopy enabled real-time monitoring of LY-11h amorphization and melt homogenization during HME. Results: ExtraTrees and Bagging models showed promising predictive performance for key filament properties, and PAT-stage validation confirmed strong agreement with experimental values. The 15 DoE-designed ASD filaments were successfully fabricated into FDM-printed tablets with reproducible geometry. Equilibrium-solubility and in vitro dissolution studies demonstrated enhanced intestinal-pH solubility and reproducible pH-responsive release. Conclusions: Collectively, these findings establish a technological proof of concept for the manufacture of LY-11h dosage forms with adjustable formulation and geometric attributes. Further in vivo pharmacokinetic studies are required to determine whether these manufacturing capabilities translate into predictable dose–exposure relationships and individualized dose control. Full article

Large integrated station–bridge buildings combine track-bearing members, station floors, transfer structures, columns, and urban-rail facilities within a single coupled structural system. For such buildings, refined train–track–station dynamic simulations can reproduce train-induced vibration, but repeated time-history analysis remains costly when many operating conditions must be screened. This study develops a simulation-based response-database framework for multi-horizon forecasting of front-end carbody vertical acceleration (FCVA), defined here as the vertical acceleration at the front-end floor evaluation point of the leading carbody, in the integrated station–bridge building of Yichang North Railway Station. The project-specific database contains 700 operating cases constructed from 100 Latin-hypercube-sampled combinations of a dimensionless track-spectrum amplitude multiplier (TSA), structural damping ratio (DR), and track-spectrum initial moving position (TSIP), each evaluated at seven train speeds. With a sampling interval of 0.002 s, supervised samples were constructed using a 200-point historical window, and prediction horizons from 20 to 300 steps (0.04–0.60 s) were evaluated under a case-level split. Classical regression, tree ensembles, a multilayer perceptron, recurrent networks, a temporal convolutional network, and a Transformer were compared after automated hyperparameter selection. For the 20-step task, Extra Trees achieved the best performance, with a root mean squared error (RMSE) of 0.00336 m/s² and $R^2$ = 0.9958. In the independently refitted reference-fixed horizon experiment, Extra Trees retained $R^2$ = 0.9526 at the 300-step horizon, while the temporal convolutional network (TCN) RMSE increased from 0.00394 to 0.01502 m/s². The results show that the response database preserves exploitable short- to medium-range dynamic continuity, although phase drift and peak-timing uncertainty increase as the forecast horizon becomes longer. Parameter analysis indicates that train speed dominates both response energy and forecast error, whereas TSA mainly affects amplitude-related response metrics. On a common central processing unit (CPU) platform, the saved Extra Trees model processed 10,000 held-out windows in $0.1404\pm 0.0008$ s. The proposed method provides a computationally efficient response-screening and post-processing layer for design-stage assessment and operating-scenario comparison within the modeled parameter domain, complementing rather than replacing refined dynamic simulation and field validation. Full article

Leakage identification in water distribution networks must estimate leak locations and magnitudes under demand fluctuations, similar adjacent-node responses, and superimposed multiple-leak signals. This study combines physics-based joint inversion with ExtraTrees candidate re-ranking. Using an EPANET model of the Hanoi network, emitter candidates were placed at pipe midpoints, and scenarios were generated across demand periods, global demand factors, and four regional demand fluctuations. Sixteen pressure residuals and 22 flow residuals formed the observation vector. The physical stage enumerated zero- to three-leak combinations and used bounded least squares to estimate leakage flows and demand corrections; standardized pressure-flow residuals and penalty terms produced the candidate pool. ExtraTrees then re-ranked candidates using candidate structure, physical scores, flow statistics, and operating-period features. In 2000 independent blind-test scenarios, complete localization accuracy, leak-number identification accuracy, and total leakage-flow MAE were 90.55%, 97.10%, and 0.841 L/s; for 1500 leakage scenarios, they were 87.53%, 96.27%, and 1.117 L/s. Re-ranking increased leakage-scenario complete localization from 76.13% to 87.53%, while final candidate-pool recall reached 99.15%. Robustness tests involving measurement noise, hydraulic-model mismatch, sensor density, computational time, and Net1 indicated that the method improves candidate discrimination under simulation, although roughness bias and weak multiple-leak signals remain challenging. Full article

Snow density influences Arctic climate, ecosystems, and surface energy exchange, yet spatially continuous observations remain limited. This study constructed an ERA5-supervised snow-density dataset for 60–90° N by collocating Advanced Microwave Scanning Radiometer 2 (AMSR-2) Level-1R brightness temperatures with ECMWF Reanalysis v5 (ERA5) snow density, Soil Moisture Active Passive (SMAP) surface roughness, and auxiliary variables. Ten models were evaluated using 29 observation days spanning September 2022–February 2023 under a chronological training–validation–test split. Extra Trees achieved the best overall performance, with a root mean square error of 18.54 kg m⁻³ and an R² of 0.87, while the bidirectional gated recurrent unit (BiGRU) was the strongest deep-learning model. Feature-attribution and ablation analyses showed that microwave brightness temperatures contained predictive information, although geographic and auxiliary variables also contributed substantially. The evaluated models could reproduce ERA5-referenced Arctic snow-density patterns, but their performance partly reflected regional information. Moreover, ERA5 showed limited consistency with station-based Northern Hemisphere Snow Water Equivalent estimates. Consequently, the reported metrics quantify agreement with ERA5 rather than accuracy against independently observed snow density. Temporally coincident and spatially independent field validation remains necessary in the future. Full article