Predicting Ketamine Exposure in Pediatric Status Epilepticus: A Pharmacokinetic-Machine Learning Approach.

In the current paper we assess different machine learning (ML) models and hybrid geostatistical methods in the prediction of soil pH using digital elevation model derivates (environmental covariates) and co-located soil parameters (soil covariates). The study was located in the area of Grevena, Greece, where 266 disturbed soil samples were collected from randomly selected locations and analyzed in the laboratory of the Soil and Water Resources Institute. The different models that were assessed were random forests (RF), random forests kriging (RFK), gradient boosting (GB), gradient boosting kriging (GBK), neural networks (NN), and neural networks kriging (NNK) and finally, multiple linear regression (MLR), ordinary kriging (OK), and regression kriging (RK) that although they are not ML models, they were used for comparison reasons. Both the GB and RF models presented the best results in the study, with NN a close second. The introduction of OK to the ML models’ residuals did not have a major impact. Classical geostatistical or hybrid geostatistical methods without ML (OK, MLR, and RK) exhibited worse prediction accuracy compared to the models that included ML. Furthermore, different implementations (methods and packages) of the same ML models were also assessed. Regarding RF and GB, the different implementations that were applied (ranger-ranger, randomForest-rf, xgboost-xgbTree, xgboost-xgbDART) led to similar results, whereas in NN, the differences between the implementations used (nnet-nnet and nnet-avNNet) were more distinct. Finally, ML models tuned through a random search optimization method were compared with the same ML models with their default values. The results showed that the predictions were improved by the optimization process only where the ML algorithms demanded a large number of hyperparameters that needed tuning and there was a significant difference between the default values and the optimized ones, like in the case of GB and NN, but not in RF. In general, the current study concluded that although RF and GB presented approximately the same prediction accuracy, RF had more consistent results, regardless of different packages, different hyperparameter selection methods, or even the inclusion of OK in the ML models’ residuals.

<b>Background:</b><br /> Pediatric status epilepticus (SE) is a neurological emergency associated with significant morbidity. The <b>Status Epilepticus Pediatric Severity Score (STEPSS)</b> offers rapid bedside prognostication, but its performance relative to machine learning (ML) models has not been well studied in resource-limited settings.<br /> <b>Methods:</b><br /> We prospectively enrolled <b>100 children</b> with SE (median age 3.4 years; 52 % male) admitted to a tertiary center in South India (2023–2024). Clinical features, investigations, and outcomes were recorded. Functional outcome was assessed at discharge using the <b>Pediatric Overall Performance Category (POPC)</b>, with unfavorable outcome defined as POPC ≥3. Prognostic accuracy of STEPSS and three ML models—<b>Logistic Regression (LR), Random Forest (RF), and XGBoost</b>—was evaluated using sensitivity, specificity, predictive values, accuracy, and area under the ROC curve (AUC).<br /> <b>Results:</b><br /> Overall, <b>30 % of children had unfavorable POPC outcomes</b>. At presentation, <b>40 % had altered consciousness</b> and <b>45 % had high-risk seizure types</b>. <b>STEPSS ≥3</b> was associated with ICU admission, mechanical ventilation, and poor outcome. STEPSS demonstrated good discrimination (<b>sensitivity 79 %, specificity 78 %, NPV 84 %, AUC 0.83</b>). Additional predictors of poor outcome included <b>low SpO₂, hyperglycemia, CT abnormalities (present in 41 % of those imaged), and delay to first AED</b> EEG was performed in 53 % of patients, with abnormalities in 38 % <b>.</b>Among ML models, LR achieved performance similar to STEPSS (AUC 0.82), while RF (AUC 0.88) and XGBoost (AUC 0.91) outperformed it, with XGBoost achieving the highest accuracy (90 %) and the fewest misclassifications. Feature importance analysis highlighted CT abnormalities, treatment delay, blood glucose, and SpO₂ as dominant predictors, with STEPSS also contributing significantly.<br /> <b>Conclusion:</b><br /> STEPSS remains a practical and reliable bedside triage tool in pediatric SE, particularly in low-resource emergency settings. However, integrating additional clinical indicators into ensemble ML models, especially XGBoost, provides superior prognostic accuracy. A <b>hybrid strategy—STEPSS combined with ML augmentation—offers both interpretability and precision</b>, supporting early risk stratification and informed critical care decisions.

Machine learning approaches for formation matrix volume prediction from well logs: Insights and lessons learned

647 Background: The primary treatment of most mNSTC is chemotherapy followed by surgery if the residual disease (RD) is >1 cm. However, conventional imaging lacks the specificity to characterize the tissue, often leading to overtreatment. This study hypothesizes that integrating CT-driven radiomics features with plasma miR371 and miR375 will enhance the predictive accuracy of Machine Learning (ML) models to predict teratoma, viable germ cell (vGCT) and fibrosis/necrosis (F/N) in mNSTC patients with RD. Methods: 111 lesions from52 patients, including residual teratoma (n=57), F/N (n=33), vGCT (n=10), and additional seminoma (n=11) for training purposes were included, split into training (N=78) and test cohorts (N=33). Lesions were lymph nodes (n=87), lung (n=21), and brain (n=3) with a median size of 1.6 cm (Q1-Q3 interval=1.2-2.73 cm). 3D Slicer version 5.6.1 was used to segment the RD > 1 cm (short axis) and extract radiomics features. Plasma miRNA levels before resection were measured by RT-PCR. Random Forest (RF), Support Vector Machine (SVM), Gradient Boosting (GB), and CatBoost (CB) ML models were evaluated to define the operating characteristics of radiomics alone (R-only) and in combination with miR371 (371) and/or miR375 (375) levels in predicting teratoma, vGCT and F/N. Results: For predicting teratoma, the best models were RF (R+375 and R+371+375), CB (R+371+375), and GB (R+371 and R+371+375). While adding miR371 or miR375 to R-only slightly improved AUC across models, the best results were achieved with the R+375+371 dataset. CB achieved AUCs ranging from 0.94 to 0.97 in training and 0.81 to 0.93 in test sets, with its highest AUC of 0.93 (95% CI: 0.78-0.97) on the R+375+371 dataset to differentiate all three classes. Similarly, GB demonstrated strong performance, achieving its highest AUC of 0.93 (95% CI: 0.79-0.96) on the R+375+371 dataset (Table). Conclusions: Integration of plasma miR371, miR375 and radiomics improved accuracy of predicting histologies across all ML models. These methods could be used to characterize the histology of RD in mNSTC patients to better inform treatment decisions. Further refinement, including incorporation of histological findings of the primary tumor, will be reported. AUC values of different ML algorithms on training and test sets. TRAINING SET TEST SET Model ±SD R R+375 R+371 R+375+371 Model (95% CI) R R+375 R+371 R+375+371 RF 0.93±0.05 0.95±0.04 0.95±0.03 0.96±0.04 RF 0.8(0.59-0.89) 0.85(0.72-0.93) 0.87(0.76-0.95) 0.91(0.78-0.95) SVM 0.84±0.06 0.84±0.09 0.89±0.11 0.89±0.09 SVM 0.72(0.54-0.80) 0.74(0.56-0.82) 0.83(0.69-0.92) 0.84(0.76-0.94) GB 0.94±0.04 0.91±0.08 0.95±0.05 0.97±0.03 GB 0.84(0.61-0.96) 0.89(0.77-0.97) 0.89(0.79-0.96) 0.93(0.79-0.96) CB 0.95±0.03 0.94±0.03 0.94±0.04 0.97±0.03 CB 0.81(0.6-0.93) 0.86(0.73-0.94) 0.89(0.78-0.97) 0.93(0.78-0.97)

BackgroundHypertension is the most common modifiable risk factor for cardiovascular diseases in South Asia. Machine learning (ML) models have been shown to outperform clinical risk predictions compared to statistical methods, but studies using ML to predict hypertension at the population level are lacking. This study used ML approaches in a dataset of three South Asian countries to predict hypertension and its associated factors and compared the model's performances.MethodsWe conducted a retrospective study using ML analyses to detect hypertension using population-based surveys. We created a single dataset by harmonizing individual-level data from the most recent nationally representative Demographic and Health Survey in Bangladesh, Nepal, and India. The variables included blood pressure (BP), sociodemographic and economic factors, height, weight, hemoglobin, and random blood glucose. Hypertension was defined based on JNC-7 criteria. We applied six common ML-based classifiers: decision tree (DT), random forest (RF), gradient boosting machine (GBM), extreme gradient boosting (XGBoost), logistic regression (LR), and linear discriminant analysis (LDA) to predict hypertension and its risk factors.ResultsOf the 8,18,603 participants, 82,748 (10.11%) had hypertension. ML models showed that significant factors for hypertension were age and BMI. Ever measured BP, education, taking medicine to lower BP, and doctor's perception of high BP was also significant but comparatively lower than age and BMI. XGBoost, GBM, LR, and LDA showed the highest accuracy score of 90%, RF and DT achieved 89 and 83%, respectively, to predict hypertension. DT achieved the precision value of 91%, and the rest performed with 90%. XGBoost, GBM, LR, and LDA achieved a recall value of 100%, RF scored 99%, and DT scored 90%. In F1-score, XGBoost, GBM, LR, and LDA scored 95%, while RF scored 94%, and DT scored 90%. All the algorithms performed with good and small log loss values <6%.ConclusionML models performed well to predict hypertension and its associated factors in South Asians. When employed on an open-source platform, these models are scalable to millions of people and might help individuals self-screen for hypertension at an early stage. Future studies incorporating biochemical markers are needed to improve the ML algorithms and evaluate them in real life.

Modelling of shallow landslides with machine learning algorithms

Hospital acquired thrombocytopenia (HAT) is a common hematological complication after surgery. This research aimed to develop and compare the performance of seven machine learning (ML) algorithms for predicting patients that are at risk of HAT after surgery. We conducted a retrospective cohort study which enrolled adult patients transferred to the intensive care unit (ICU) after surgery in West China Hospital of Sichuan University from January 2016 to December 2018. All subjects were randomly divided into a derivation set (70%) and test set (30%). ten-fold cross-validation was used to estimate the hyperparameters of ML algorithms during the training process in the derivation set. After ML models were developed, the sensitivity, specificity, area under the curve (AUC), and net benefit (decision analysis curve, DCA) were calculated to evaluate the performances of ML models in the test set. A total of 10,369 patients were included and in 1354 (13.1%) HAT occurred. The AUC of all seven ML models exceeded 0.7, the two highest were Gradient Boosting (GB) (0.834, 0.814–0.853, p < 0.001) and Random Forest (RF) (0.828, 0.807–0.848, p < 0.001). There was no difference between GB and RF (0.834 vs. 0.828, p = 0.293); however, these two were better than the remaining five models (p < 0.001). The DCA revealed that all ML models had high net benefits with a threshold probability approximately less than 0.6. In conclusion, we found that ML models constructed by multiple preoperative variables can predict HAT in patients transferred to ICU after surgery, which can improve risk stratification and guide management in clinical practice.

Improving asphalt mix design by predicting alligator cracking and longitudinal cracking based on machine learning and dimensionality reduction techniques

The permeability coefficient of soils is an essential measure for designing geotechnical construction. The aim of this paper was to select a highest performance and reliable machine learning (ML) model to predict the permeability coefficient of soil and quantify the feature importance on the predicted value of the soil permeability coefficient with aided machine learning‐based SHapley Additive exPlanations (SHAP) and Partial Dependence Plot 1D (PDP 1D). To acquire this purpose, five single ML algorithms including K‐nearest neighbors (KNN), support vector machine (SVM), light gradient boosting machine (LightGBM), random forest (RF), and gradient boosting (GB) are used to build ML models for predicting the permeability coefficient of soils. Performance criteria for ML models include the coefficient of correlation R 2 , root mean square error (RMSE), mean absolute percentage error (MAPE), and mean absolute error (MAE). The best performance and reliable single ML model for predicting the permeability coefficient of soil for the testing dataset is the gradient boosting (GB) model, which has R 2 = 0.971, RMSE = 0.199 × 10 −11 m/s, MAE = 0.161 × 10 −11 m/s, and MAPE = 0.185%. To identify and quantify the feature importance on the permeability coefficient of soil, sensitivity studies using permutation importance, SHapley Additive exPlanations (SHAP), and Partial Dependence Plot 1D (PDP 1D) are performed with the aided best performance and reliable ML model GB. Plasticity index, density &gt; water content, liquid limit, and plastic limit &gt; clay content &gt; void ratio are the order effects on the predicted value of the permeability coefficient. The plasticity index and density of soil are the first priority soil properties to measure when assessing the permeability coefficient of soil.

Buckling and ultimate load prediction models for perforated steel beams using machine learning algorithms

Optimisation and interpretation of machine and deep learning models for improved water quality management in Lake Loktak

Myocardial perfusion imaging (MPI) is an essential tool used to diagnose and manage patients with suspected or known coronary artery disease. Additionally, the General Data Protection Regulation (GDPR) represents a milestone about individuals' data security concerns. On the other hand, Machine Learning (ML) has had several applications in the most diverse knowledge areas. It is conceived as a technology with huge potential to revolutionize health care. In this context, we developed ML models to evaluate their ability to distinguish an individual's sex from MPI assessment. We used 260 polar maps (140 men/120 women) to train ML algorithms from a database of patients referred to a university hospital for clinically indicated MPI from January 2016 to December 2018. We tested 07 different ML models, namely, Classification and Regression Tree (CART), Naive Bayes (NB), K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Adaptive Boosting (AB), Random Forests (RF) and, Gradient Boosting (GB). We used a cross-validation strategy. Our work demonstrated that ML algorithms could perform well in assessing the sex of patients undergoing myocardial scintigraphy exams. All the models had accuracy greater than 82%. However, only SVM achieved 90%. KNN, RF, AB, GB had, respectively, 88, 86, 85, 83%. Accuracy standard deviation was lower in KNN, AB, and RF (0.06). SVM and RF had had the best area under the receiver operating characteristic curve (0.93), followed by GB (0.92), KNN (0.91), AB, and NB (0.9). SVM and AB achieved the best precision. Our results bring some challenges regarding the autonomy of patients who wish to keep sex information confidential and certainly add greater complexity to the debate about what data should be considered sensitive to the light of the GDPR.

Machine learning (ML) and deep learning (DL) models can play an important role when it comes to modelling complicated processes. Such capability is necessary for hydrological and climate-related applications. Generally, ML models utilize precipitation and temperature time series of a basin as input to develop a lumped rainfall-runoff model to simulate streamflow at the basin outlet. However, when it is divided into several sub-basins, Graph Neural Networks (GNN) can consider each sub-basin as a node and link them together using a connectivity matrix to account for spatial variations of hydroclimatic variables. In this study, GNN and various ML models with different types of architecture, ranging from neural networks, tree-based structure, and gradient boosting, were exploited for daily streamflow simulation over different case studies. For each case study, the basin was divided into a few sub-basins for which daily precipitation and temperature data were aggregated and used as input. For training GNN, the connection matrix of sub-basins was also used as input. Basically, 75% of historical records were utilized to train GNN and different ML models, e.g., artificial neural networks, support vector machine, decision tree, random forest, eXtreme Gradient Boosting (XGBoost), Light Gradient-Boosting Machine (LightGBM), and Category Boosting (CatBoost), while the rest was used for testing. Streamflow simulation was conducted with/without considering seasonality impact and lag times. The obtained results clearly demonstrate that considering seasonality and time lags can enhance accuracy of streamflow predictions based on Kling–Gupta efficiency (KGE). Furthermore, GNN with seasonality impact and time lags achieved promising results across different case studies with KGE&gt;0.85 for training and KGE&gt;0.59 for testing data, respectively. Among ML models, boosting models, e.g., LightGBM and XGBoost, performed slightly better than other ML models. for Finally, this comparative analysis provides valuable insights for ML/DL applications in climate change impact assessments.Acknowledgements: This research work was carried out as part of the TRANSCEND project with funding received from the European Union Horizon Europe Research and Innovation Programme under Grant Agreement No. 10108411.

An accurate total skin factor prediction for an oil well is critical for the evaluation of the inflow performance relationship, and the optimization of the appropriate stimulation treatment such as acidizing and hydraulic fracturing. Performing well testing regularly is not economically feasible, and the equations used for total skin damage may not be accurate. In this work, the goal is to build machine learning (ML) models that can predict the total skin factor in perforated wells using accessible field data. Nine distinct ML algorithms such as Gradient Boosting (GB), Adaptive Boosting (AdaBoost), Random Forest (RF), Support Vector Machines (SVMs), Decision Trees (DT), K-Nearest Neighbor (KNN), Linear Regression (LR), Stochastic Gradient Descent (SGD), and Artificial Neural Network (ANN) are meticulously developed and fine-tuned using a substantial dataset derived from 1,088 wells. The dataset encompasses 19,040 data points, thoughtfully split into two subsets: 70% (13,328 data points) for training the algorithms, and 30% (5,712 data points) for testing their predictions. The parameters used are mostly gathered during well completion and conventional well testing operations, including liquid flow rate, water cut, gas oil ratio, bottomhole flowing pressure, reservoir pressure, reservoir temperature, reservoir permeability, reservoir thickness, perforations diameter, perforations density, perforations penetration depth, well deviation, and penetrated portion of the net pay thickness. In this study, the total skin factor acquired from conventional well test analysis serves as the model's output. K-fold cross-validation and repeated random sampling validation techniques are used to assess the performance of the models against the total skin obtained from the conventional well test analysis. The K-fold cross-validation outcomes of the top-performing ML models, specifically GB, AdaBoost, RF, DT, and KNN, reveal remarkably low mean absolute percentage error values reported as 3.2%, 3.2%, 2.9%, 3.3%, and 3.8%, respectively. Additionally, the correlation coefficients (R2) for these models are notably high, with values of 0.972, 0.968, 0.975, 0.964, and 0.956, respectively. In conclusion, ML models demonstrated their ability to predict total skin factor for different reservoir fluid properties, well geometries, and completion configurations. ML models offer a more efficient, quick, and cost-effective alternative to the conventional well-testing analysis.

Agriculture is a backbone of the Indian economy and people's lives. In agriculture land, soil is the most important element on which the quality of production and efficiency depends to the maximum extent. Phosphorus (P), Nitrogen (N), Potassium (K), and the potential of hydrogen (pH) are the key nutrients in soil. An efficient crop recommender and prediction system is needed to optimize agriculture practices considering the escalating demand for more food. Traditional time‐consuming and manual farming should be replaced with a smart agriculture framework using the integration of technologies like the Internet of Things (IoT), Wireless Sensor Network (WSN), and Machine Learning (ML). This paper proposed an IoT‐WSN driven crop management system with Neuro‐ML Ensemble Model, utilizing LoRaWAN Gateway, that can be deployed in the agriculture field to collect real‐time soil parameters. In this paper for soil nutrient analysis, the author used various ML algorithms such as Naive Bayes (NB), Logistic Regression (LR), K‐Nearest Neighbor (KNN), Decision Tree (DT), Random Forest (RF), Ada Boost (AB), Gradient Boosting (GB), and Support Vector Machine (SVM) and recommending a suitable ML algorithm for the crop recommender system. For crop yield prediction, the author has developed and recommended a customized GB Algorithm with an accuracy of 98.80%, and for the fertilizer recommendation system, the author has suggested CNN‐BiGRU which outperforms other approaches like BiGRU and CNN with an average accuracy rate of 92.48%. The author presented work with respect to the Indian agriculture sector and compared ML algorithms with state‐of‐the‐art datasets available on some government websites of India, and used by other authors, with a dataset collected by the author from hardware using Raspberry Pi. For crop recommendation and forecasting, the Neuro‐ML Ensemble model employs the Neuro‐ML, which combines neural networks (NN) with the ML models. This research aspires to assist farmers in opting for suitable crops as per their environmental suitability and situation by analyzing and predicting which crops suit well to fit the parameters required to enhance crop growth like soil nutrients, soil moisture, soil pH, and rainfall, etc. The author obtained accuracy for various ML models used in the framework. For NB, LR, KNN, SVM, DT, and RF, the author obtained accuracies of 99.54%, 96.36%, 95.90%, 96.81%, 98.86%, and 99.31%, respectively, using the Kaggle dataset available as open access. Through a dataset collected by the authors, we obtained accuracies of 94.54%, 91.36%, 92.72%, 92.73%, 86.36%, and 94.54% for NB, LR, KNN, SVM, DT, and RF, respectively. The author found that Naive Bayes (NB) outperforms the other machine learning algorithms, such as KNN, SVM, LR, Decision Tree, RF, and AB, and is the best algorithm suited for crop yield.