Search Results (847)

Background: A short cervix in the second trimester significantly increases preterm birth risk, yet no reliable first-trimester prediction method exists. Current guidelines lack consensus on which women should undergo transvaginal ultrasound (TVUS) screening for cost-effective prevention. Therefore, it is vital to establish a highly accurate and economical method for use in the early stages of pregnancy to predict short cervix in mid-pregnancy. Methods: A total of 1480 pregnant women with singleton pregnancies and at least one risk factor for spontaneous preterm birth (<37 weeks) were recruited from January 2020 to December 2020 at the Shanghai First Maternity and Infant Hospital, Tongji University School of Medicine. Cervical length was assessed at 20–24 weeks of gestation, with a short cervix defined as <25 mm. Feature selection employed tree models, regularization, and recursive feature elimination (RFE). Seven machine learning models (logistic regression, linear discriminant analysis, k-nearest neighbors, support vector machine, decision tree, random forest, XGBoost) were trained to predict mid-trimester short cervix. The XGBoost model—an ensemble method leveraging sequential decision trees—was analyzed using Shapley Additive Explanation (SHAP) values to assess feature importance, revealing consistent associations between clinical predictors and outcomes that align with known clinical patterns. Results: Among 1480 participants, 376 (25.4%) developed mid-trimester short cervix. The XGBoost-based prediction model demonstrated high predictive performance in the training set (Recall = 0.838, F1 score = 0.848), test set (Recall = 0.850, F1 score = 0.910), and an independent dataset collected in January 2025 (Recall = 0.708, F1 score = 0.791), with SHAP analysis revealing pre-pregnancy BMI as the strongest predictor, followed by second-trimester pregnancy loss history, peripheral blood leukocyte count (WBC), and positive vaginal microbiological culture results (≥10⁵ CFU/mL, measured between 11⁺⁰ and 13⁺⁶ weeks). Conclusions: The XGBoost model accurately predicts mid-trimester short cervix using first-trimester clinical data, providing a 6-week window for targeted interventions before the 20–24-week gestational assessment. This early prediction could help guide timely preventive measures, potentially reducing the risk of spontaneous preterm birth (sPTB). Full article

Accurate mapping of seafloor morphological features, such as pockmarks, is essential for marine spatial planning, geological hazard assessment, and environmental monitoring. Traditional manual delineation methods are often subjective and inefficient when applied to large, high-resolution bathymetric datasets. This study presents a semi-automated workflow based on the CoMMa (Confined Morphologies Mapping) toolbox to classify pockmarks in Flensburg Fjord, Germany–Denmark. Initial detection employed the Bathymetric Position Index (BPI) with intentionally permissive parameters to ensure high recall of morphologically diverse features. Morphometric descriptors were then extracted and used to train a Random Forest classifier, enabling noise reduction and refinement of overinclusive delineations. Validation against expert-derived mappings showed that the model achieved an overall classification accuracy of 86.16%, demonstrating strong performance across the validation area. These findings highlight how integrating a GIS-based geomorphometry toolbox with machine learning yields a reproducible, objective, and scalable approach to seabed mapping, supporting decision-making processes and advancing standardized methodologies in marine geomorphology. Full article

Leaf area index (LAI) is a crucial parameter reflecting the crown structure of the grassland. Accurately obtaining LAI is of great significance for estimating carbon sinks in grassland ecosystems. However, spectral noise interference and pronounced spatial heterogeneity within vegetation canopies constitute significant impediments to achieving high-precision LAI retrieval. This study used hyperspectral sensor mounted on an unmanned aerial vehicle (UAV) to estimate LAI in a typical grassland, Hulun Lake Basin. Multiple machine learning (ML) models were constructed to reveal a relationship between hyperspectral data and grassland LAI using two input datasets, namely spectral transformations and vegetation indices (VIs), while SHAP (SHapley Additive ExPlanation) interpretability analysis was further employed to identify high-contribution features in the ML models. The analysis revealed that grassland LAI has good correlations with the original spectrum at 550 nm and 750 nm–1000 nm, first and second derivatives at 506 nm–574 nm, 649 nm–784 nm, and vegetation indices including the triangular vegetation index (TVI), enhanced vegetation index 2 (EVI2), and soil-adjusted vegetation index (SAVI). In the models using spectral transformations and VIs, the random forest (RF) models outperformed other models (testing R² = 0.89/0.88, RMSE = 0.20/0.21, and RRMSE = 27.34%/28.98%). The prediction error of the random forest model exhibited a positive correlation with measured LAI magnitude but demonstrated an inverse relationship with quadrat-level species richness, quantified by Margalef’s richness index (MRI). We also found that at the quadrat level, the spectral response curve pattern is influenced by attributes within the quadrat, like dominant species and vegetation cover, and that LAI has positive relationship with quadrat vegetation cover. The LAI inversion results in this study were also compared to main LAI products, showing a good correlation (r = 0.71). This study successfully established a high-fidelity inversion framework for hyperspectral-derived LAI estimation in mid-to-high latitude grasslands of the Hulun Lake Basin, supporting the spatial refinement of continental-scale carbon sink models at a regional scale. Full article

This paper addresses the problem of recognising and classifying hand movements to control an upper-limb exoskeleton. To solve this problem, a multisensory system based on the fusion of data from electromyography (EMG) sensors, inertial measurement units (IMUs), and virtual reality (VR) trackers is proposed, which provides highly accurate detection of users’ movements. Signal preprocessing (noise filtering, segmentation, normalisation) and feature extraction were performed to generate input data for regression and classification models. Various machine learning algorithms are used to recognise motor activity, ranging from classical algorithms (logistic regression, k-nearest neighbors, decision trees) and ensemble methods (random forest, AdaBoost, eXtreme Gradient Boosting, stacking, voting) to deep neural networks, including convolutional neural networks (CNNs), gated recurrent units (GRUs), and transformers. The algorithm for integrating machine learning models into the exoskeleton control system is considered. In experiments aimed at abandoning proprietary tracking systems (VR trackers), absolute position regression was performed using data from IMU sensors with 14 regression algorithms: The random forest ensemble provided the best accuracy (mean absolute error = 0.0022 metres). The task of classifying activity categories out of nine types is considered below. Ablation analysis showed that IMU and VR trackers produce a sufficient informative minimum, while adding EMG also introduces noise, which degrades the performance of simpler models but is successfully compensated for by deep networks. In the classification task using all signals, the maximum result (99.2%) was obtained on Transformer; the fully connected neural network generated slightly worse results (98.4%). When using only IMU data, fully connected neural network, Transformer, and CNN–GRU networks provide 100% accuracy. Experimental results confirm the effectiveness of the proposed architectures for motor activity classification, as well as the use of a multi-sensor approach that allows one to compensate for the limitations of individual types of sensors. The obtained results make it possible to continue research in this direction towards the creation of control systems for upper exoskeletons, including those used in rehabilitation and virtual simulation systems. Full article

Knee joint stiffness occurs and severely limits its range of motion (ROM) after facture around the knee. During mobility training, knee joints need to be flexed to the maximum angle position (maxAP) that can induce pain at an appropriate level in order to pull apart intra-articular adhesive structures while avoiding secondary injuries. However, the maxAP varies with training and is mostly determined by the pain level of patients. In this study, the feasibility of utilizing electromyogram (EMG) activities to detect maxAP was investigated. Specifically, the maxAP detection was converted into a binary classification between pain level three of the numerical rating scales (pain) and below (painless) according to clinical requirements. Firstly, 12 post-fracture patients with knee joint stiffness participated in Experiment I, with a therapist performing routine mobility training and EMG signals being recorded from knee flexors and extensors. The results showed that the extracted EMG features were significantly different between the pain and painless states. Then, the maxAP estimation performance was tested on a knee rehabilitation robot in Experiment II, with another seven patients being involved. The support vector machine and random forest models were used to classify between pain and painless states and obtained a mean accuracy of 87.90% ± 4.55% and 89.10% ± 4.39%, respectively, leading to an average estimation bias of 6.5° ± 5.1° and 4.5° ± 3.5°. These results indicated that the pain-induced EMG can be used to accurately classify pain states for the maxAP estimation in post-fracture mobility training, which can potentially facilitate the application of robotic techniques in fracture rehabilitation. Full article

The compressive strength (Fc) of cement mortar (CM) is a key parameter in ensuring the mechanical reliability and durability of cement-based materials. Traditional testing methods are labor-intensive, time-consuming, and often lack predictive flexibility. With the increasing adoption of machine learning (ML) in civil engineering, data-driven approaches offer a rapid, cost-effective alternative for forecasting material properties. This study investigates a wide range of supervised linear and nonlinear ML regression models to predict the Fc of CM. The evaluated models include linear regression, ridge regression, lasso regression, decision trees, random forests, gradient boosting, k-nearest neighbors (KNN), and twelve neural network (NN) architectures, developed by combining different optimizers (L-BFGS, Adam, and SGD) with activation functions (tanh, relu, logistic, and identity). Model performance was assessed using the root mean squared error (RMSE), coefficient of determination (R²), and mean absolute error (MAE). Among all models, NN_tanh_lbfgs achieved the best results, with an almost perfect fit in training (R² = 0.9999, RMSE = 0.0083, MAE = 0.0063) and excellent generalization in testing (R² = 0.9946, RMSE = 1.5032, MAE = 1.2545). NN_logistic_lbfgs, gradient boosting, and NN_relu_lbfgs also exhibited high predictive accuracy and robustness. The SHAP analysis revealed that curing age and nano silica/cement ratio (NS/C) positively influence Fc, while porosity has the strongest negative impact. The main novelty of this study lies in the systematic tuning of neural networks via distinct optimizer–activation combinations, and the integration of SHAP for interpretability—bridging the gap between predictive performance and explainability in cementitious materials research. These results confirm the NN_tanh_lbfgs as a highly reliable model for estimating Fc in CM, offering a robust, interpretable, and scalable solution for data-driven strength prediction. Full article

Extreme sea surface temperature (SST) events, such as marine heatwaves (MHWs) and marine cold spells (MCSs), severely affect warm water coral reefs. However, further study is required on their historical and future spatiotemporal patterns, driving mechanisms, and impacts in coral reef regions. This study analyzed the spatiotemporal patterns in MHWs/MCSs for the periods 1982–2022 and 2023–2070 using ten indices based on OISSTv2.1 and CMIP6 data, respectively, identified key MHW drivers via four machine learning methods (Random Forest, Extreme Gradient Boosting, Light Gradient Boosting Machine, and categorical boosting) and SHAP values (Shapley Additive Explanations), and then examined their relationship with coral coverage across ten global marine regions. Our results revealed that (1) MHWs are not only increasing in their average intensity but also becoming more extreme, while MCSs have declined. More MHW days are observed in regions like the Red Sea, the Persian Gulf, and the South Pacific Islands, with increases of up to 28 days per decade. (2) Higher-latitude coral reefs are experiencing more severe MHWs than equatorial regions, with up to 1.24 times more MHW days, emphasizing the urgent need to protect coral refuges. (3) MHWs are projected to occur nearly year-round by 2070 under scenario SSP5–8.5. The area ratio of MHWs to MCSs is expected to rise sharply from 2040 onward, reaching approximately 100-fold under the SSP2–4.5 scenario and 196-fold under the SSP5–8.5 scenario, particularly in the Marshall Islands and Caribbean Sea regions. (4) The coefficient of variation (CV) of annual temperature, annual ocean heat content, and monthly temperature were the top three factors driving MHW intensity. We emphasize that future MHW predictions should focus more on the CV of forecasting indicators rather than just the climate means. (5) Coral coverage exhibited post-mortality processes following MHWs, showing a strong negative correlation (r = −0.54, p < 0.01) with MHWs while demonstrating a significant positive correlation (r = 0.6, p < 0.01) with MCSs. Our research underscores the sustained efforts to protect and restore coral reefs amid escalating climate-induced stressors. Full article

Mangrove forests are among the world’s most vital coastal ecosystems. Mapping mangrove cover from local to global scales using spectral data and machine learning models is a well-established method. While non-spectral contextual datasets (spatial features) have also been incorporated into such models, the contribution of these additional features to improving mangrove mapping remains underexplored. Using the Mesoamerican Reef Region as a case study, we evaluate the effectiveness of incorporating spatial features in binary mangrove classification to enhance mapping accuracy. We compared an aspatial model that includes only spectral data with three spatial models: two included features such as geographic coordinates, elevation, and proximity to coastlines and streams, while the third integrated a geostatistical approach using Inverse Distance Weighted (IDW) interpolation. Spectral inputs included bands and indices derived from Sentinel-1 and Sentinel-2, and all models were implemented using the Random Forest algorithm in Google Earth Engine. Results show that spatial features reduced omission errors without increasing commission errors, enhancing the model’s ability to capture spatial variability. Models using geographic coordinates and elevation performed comparably to those with additional environmental variables, with storm frequency and distance to streams emerging as important predictors in the Mesoamerican Reef region. In contrast, the IDW-based model underperformed, likely due to overfitting and limited representation of local spectral variation. Spatial analyses show that models incorporating spatial features produced more continuous mangrove patches and removed some false positives in non-mangrove areas. These findings highlight the value of spatial features in improving classification accuracy, especially in regions with ecologically diverse mangroves across varied environments. By integrating spatial context, these models support more accurate, locally relevant mangrove maps that are essential for effective conservation and management. Full article

To examine the integrated influence of urban spatial form on the urban heat island (UHI) effect, this study selects the area within Beijing’s Fifth Ring Road as a case study. A multiscale grid system is established to quantify fourteen two- and three-dimensional morphological indicators. A Random Forest algorithm is employed to assess the relative importance of each factor. The optimal analytical scale for each key variable is then identified, and its nonlinear relationship with land surface temperature (LST) is analyzed at that scale. The main findings are as follows: (1) The Random Forest model achieves the highest predictive accuracy at a 600 m scale, significantly outperforming traditional linear models by effectively addressing multicollinearity. This suggests that machine learning offers robust technical support for UHI research. (2) Form variables exhibit distinct scale dependencies. Two-dimensional indicators dominate at medium to large scales, while three-dimensional indicators are more influential at smaller scales. Specifically, the mean building height is most significant at the 150 m scale, the standard deviation of building height at 300 m, and the impervious surface fraction at 600–1200 m. (3) Strong nonlinear effects are identified. The bare soil fraction below 0.12 intensifies surface warming; the water body fraction between 0.20 and 0.35 provides the strongest cooling; plant coverage offers maximum cooling between 0.25 and 0.45; building density cools below 0.3 buildings/hm² but contributes to warming beyond this threshold; building coverage ratio generates the greatest warming between 0.08 and 0.32; height variability provides optimal cooling between 8 m and 40 m; and mean building height shows a positive correlation with LST below 6 m but a negative one above that height. Full article

The study is devoted to the construction of an avalanche susceptibility map based on ensemble machine learning algorithms (random forest, XGBoost, LightGBM, gradient boosting machines, AdaBoost, NGBoost) for the conditions of the East Kazakhstan region. To train these models, data were collected on avalanche path profiles, meteorological conditions, and historical avalanche events. The quality of the trained machine learning models was assessed using metrics such as accuracy, precision, true positive rate (recall), and F1-score. The obtained metrics indicated that the trained machine learning models achieved reasonably accurate forecasting performance (forecast accuracy from 67% to 73.8%). ROC curves were also constructed for each obtained model for evaluation. The resulting AUCs for these ROC curves showed acceptable levels (from 0.57 to 0.73), which also indicated that the presented models could be used to predict avalanche danger. In addition, for each machine learning model, we determined the importance of the indicators used to predict avalanche danger. Analysis of the importance of the indicators showed that the most significant indicators were meteorological data, namely temperature and snow cover level in avalanche paths. Among the indicators that characterized the avalanche paths’ profiles, the most important were the minimum and maximum slope elevations. Thus, within the framework of this study, a highly accurate model was built using geospatial and meteorological data that allows identifying potentially dangerous slope areas. These results can support territorial planning, the design of protective infrastructure, and the development of early warning systems to mitigate avalanche risks. Full article

Acoustic sound-based condition monitoring (ASCM) systems, which typically utilize machine learning algorithms on established audio features, have demonstrated effectiveness under controlled conditions. However, their application in real-world industrial environments presents significant challenges due to complex and variable soundscapes with high noise and limited fault data. The presence of random interfering sounds and variability in operating conditions can lead to lower performance and high false-positive rates. To overcome these limitations, we propose a fault detection method that leverages the underlying physical characteristics of the sound signals. By investigating the components of the acoustic signal, we found that fault-related sounds can be modeled as exponentially decaying oscillations. This insight allows for the development of a physically based signal model, setting our approach apart from purely data-driven methods. Using this model, we developed a robust detection method based on a Generalized Likelihood Ratio Test (GLRT). The effectiveness of this approach was validated using both synthetic and real-world data from a steel industry facility. Our results demonstrate that the proposed model-based approach provides superior performance compared to standard audio features, particularly in high-noise conditions. On real-world data, the GLRT-based approach outperformed all audio features, as clearly shown by the Receiver Operating Characteristic (ROC) analysis. Specifically, the Partial Area Under the Curve (pAUC) of the GLRT is more than twice that of the best-performing audio feature, demonstrating good detection at significantly lower-false-positive rates compared to audio features. Furthermore, simulations showed that our method maintains robust detection down to a Signal-to-Noise Ratio (SNR) of −13 $\mathrm{dB}$, significantly outperforming audio feature-based detection, which was limited to approximately −10 $\mathrm{dB}$. The physically informed nature of our model not only provides a more reliable and robust solution but also enables the method to be generalized to other industrial scenarios with similar fault properties, offering broader applicability for reliable acoustic condition monitoring. Full article

With accelerating urbanization, the development of rail transit systems—particularly subways—has become a key strategy for alleviating urban traffic congestion. However, existing studies on subway station site selection often lack a spatially continuous evaluation of site suitability across the entire study area. This may lead to a disconnect between planning and actual demand, resulting in issues such as “overbuilt infrastructure” or the “island effect.” To address this issue, this study selects Kunming City, China, as the study area, employs the K-means++ algorithm to cluster existing subway stations based on passenger flow, integrates multi-source spatial data, applies a random forest algorithm for optimal positive sample selection and driving factor identification, and subsequently uses a LightGBM-SHAP explainable machine learning framework to develop a predictive model for station location based on mathematical modeling. The main findings of the study are as follows: (1) Using the random forest model, 20 key drivers influencing site selection were identified. SHAP analysis revealed that the top five contributing factors were connectivity, nighttime lighting, road network density, transportation service, and residence density. Among these, transportation-related factors accounted for three out of five and emerged as the primary determinants of subway station site selection. (2) The site selection prediction model exhibited strong performance, achieving an R² value of 0.95 on the test set and an average R² of 0.79 during spatial 5-fold cross-validation, indicating high model reliability. The spatial distribution of predicted suitability indicated that the core urban area within the Second Ring Road exhibited the highest suitability, with suitability gradually declining toward the periphery. High-suitability areas outside the Third Ring Road in suburban regions were primarily aligned along existing subway lines. (3) The cumulative predicted probability within a 300 m buffer zone around each station was positively correlated with passenger flow levels. Overlaying the predicted results with current station locations revealed strong spatial consistency, indicating that the model outputs closely align with the actual spatial layout and passenger usage intensity of existing stations. These findings provide valuable decision-making support for optimizing subway station layouts and planning future transportation infrastructure, offering both theoretical and practical significance for data-driven site selection. Full article

Archaeological research fundamentally relies on detecting features to uncover hidden historical information. Airborne (aerial) LiDAR technology has significantly advanced this field by providing high-resolution 3D terrain maps that enable the identification of ancient structures and landscapes with improved accuracy and efficiency. This technical note comprehensively reviews 45 recent studies to critically examine the integration of Machine Learning (ML) and Deep Learning (DL) techniques, particularly Convolutional Neural Networks (CNNs), with airborne LiDAR derivatives for automated archaeological feature detection. The review highlights the transformative potential of these approaches, revealing their capability to automate feature detection and classification, thus enhancing efficiency and accuracy in archaeological research. CNN-based methods, employed in 32 of the reviewed studies, consistently demonstrate high accuracy across diverse archaeological features. For example, ancient city walls were delineated with 94.12% precision using U-Net, Maya settlements with 95% accuracy using VGG-19, and with an IoU of around 80% using YOLOv8, and shipwrecks with a 92% F1-score using YOLOv3 aided by transfer learning. Furthermore, traditional ML techniques like random forest proved effective in tasks such as identifying burial mounds with 96% accuracy and ancient canals. Despite these significant advancements, the application of ML/DL in archaeology faces critical challenges, including the scarcity of large, labeled archaeological datasets, the prevalence of false positives due to morphological similarities with natural or modern features, and the lack of standardized evaluation metrics across studies. This note underscores the transformative potential of LiDAR and ML/DL integration and emphasizes the crucial need for continued interdisciplinary collaboration to address these limitations and advance the preservation of cultural heritage. Full article

Background/Objectives: ISUP grade group discordance between prostate biopsy and radical prostatectomy (RP) impacts treatment decisions in over a third (~25–40%) of prostate cancer (PCa) patients. We aimed to identify ISUP grade migration predictors and assess the impact of preoperative imaging (MRI) in a contemporary Romanian PCa cohort. Methods: We retrospectively analyzed 142 PCa patients undergoing RP following biopsy between January 2021 and December 2024 at Pius Brinzeu County Hospital, Timișoara: 90 without and 52 with preoperative MRI. Clinical parameters, MRI findings (PI-RADS), and biopsy characteristics were evaluated. Machine learning models (gradient boosting, random forest) were developed with SHAP analysis for interpretability. Results: Grade migration occurred in 69/142 patients (48.6%): upstaging in 55 (38.7%) and downstaging in 14 (9.9%). In the non-MRI cohort, 37/90 (41.1%) were upstaged and 9/90 (10.0%) were downstaged, versus 18/52 (34.6%) upstaged and 5/52 (9.6%) downstaged in the MRI cohort. The MRI group showed a 6.5% absolute reduction in upstaging (34.6% vs. 41.1%), a promising non-significant trend (p = 0.469) that requires further investigation. Grade 1 patients showed the highest upstaging (69.4%), while Grades 3–4 showed the highest downstaging (11/43, 25.6%). PI-RADS 4 lesions had the highest upstaging (43.5%). PSA density > 0.20 ng/mL² emerged as the strongest predictor. Gradient boosting achieved superior performance (AUC = 0.812) versus logistic regression (AUC = 0.721), representing a 13% improvement in discrimination. SHAP analysis revealed PSA density as the most influential (importance: 0.287). Grade migration associated with adverse pathology: extracapsular extension (52.7% vs. 28.7%, p = 0.008) and positive margins (38.2% vs. 21.8%, p = 0.045). Conclusions: ISUP grade migration affects 48.6% of Romanian patients, with 38.7% upstaged and 9.9% downstaged. The 69.4% upstaging in Grade 1 patients emphasizes the need for enhanced risk stratification tools, while 10% downstaging suggests potential overtreatment. Machine learning with SHAP analysis provides superior predictive performance (13% AUC improvement) while offering clinically interpretable risk assessments. PSA density dominates risk assessment, while PI-RADS 4 lesions warrant closer scrutiny than previously recognized. Full article

This paper explores emotional analysis in Hebrew texts, focusing on improving machine learning techniques for depression detection by integrating psychological feature lexicons. Hebrew’s complex morphology makes emotional analysis challenging, and this study seeks to address that by combining traditional machine learning methods with sentiment lexicons. The dataset consists of over 350,000 posts from 25,000 users on the health-focused social network “Camoni” from 2010 to 2021. Various machine learning models—SVM, Random Forest, Logistic Regression, and Multi-Layer Perceptron—were used, alongside ensemble techniques like Bagging, Boosting, and Stacking. TF-IDF was applied for feature selection, with word and character n-grams, and pre-processing steps like punctuation removal, stop word elimination, and lemmatization were performed to handle Hebrew’s linguistic complexity. The models were enriched with sentiment lexicons curated by professional psychologists. The study demonstrates that integrating sentiment lexicons significantly improves classification accuracy. Specific lexicons—such as those for negative and positive emojis, hostile words, anxiety words, and no-trust words—were particularly effective in enhancing model performance. Our best model classified depression with an accuracy of 84.1%. These findings offer insights into depression detection, suggesting that practitioners in mental health and social work can improve their machine learning models for detecting depression in online discourse by incorporating emotion-based lexicons. The societal impact of this work lies in its potential to improve the detection of depression in online Hebrew discourse, offering more accurate and efficient methods for mental health interventions in online communities. Full article