Model Performed Research Articles

Retransmission performance in a stochastic geometric cellular network model

Suppose sender–receiver transmission links in a downlink network at a given data rate are subject to fading, path loss, and inter-cell interference, and that transmissions either pass, suffer loss, or incur retransmission delay. We introduce a method to obtain the average activity level of the system required for handling the buffered work and from this derive the resulting coverage probability and key performance measures. The technique involves a family of stationary buffer distributions which is used to solve iteratively a nonlinear balance equation for the unknown busy-link probability and then identify throughput, loss probability, and delay. The results allow for a straightforward numerical investigation of performance indicators, are in special cases explicit and may be easily used to study the trade-off between reliability, latency, and data rate.

Nomogram based on immune-inflammatory indicators and age-adjusted charlson comorbidity index score to predict prognosis of postoperative parotid gland carcinoma patients

BackgroundParotid gland carcinoma (PGC) is a rare malignant tumor. The purpose of this study was to investigate the role of immune-inflammatory-nutrition indicators and age-adjusted Charlson comorbidity index score (ACCI) of PGC and develop the nomogram model for predicting prognosis.MethodAll patients diagnosed with PGC in two tertiary hospitals, treated with surgical resection, from March 2012 to June 2018 were obtained. Potential prognostic factors were identified by univariate and multivariate Cox regression analyses. The nomogram models were established based on these identified independent prognostic factors. The performance of the developed prognostic model was estimated by related indexes and plots.ResultThe study population consisted of 344 patients with PGC who underwent surgical resection, 285 patients without smoking (82.8%), and 225 patients (65.4%) with mucoepidermoid carcinoma, with a median age of 50.0 years. American Joint Committee on Cancer (AJCC) stage (p < 0.001), pathology (p = 0.019), tumor location (p < 0.001), extranodal extension (ENE) (p < 0.001), systemic immune-inflammation index (SII) (p = 0.004), prognostic nutrition index (PNI) (p = 0.003), ACCI (p < 0.001), and Glasgow prognostic Score (GPS) (p = 0.001) were independent indicators for disease free survival (DFS). Additionally, the independent prognostic factors for overall survival (OS) including AJCC stage (p = 0.015), pathology (p = 0.004), tumor location (p < 0.001), perineural invasion (p = 0.009), ENE (p < 0.001), systemic immune-inflammation index (SII) (p = 0.001), PNI (p = 0.001), ACCI (p = 0.003), and GPS (p = 0.033). The nomogram models for predicting DFS and OS in PGC patients were generated based on these independent risk factors. All nomogram models show good discriminative capability with area under curves (AUCs) over 0.8 (DFS 0.802, and OS 0.825, respectively). Decision curve analysis (DCA), integrated discrimination improvement (IDI), and net reclassification index (NRI) show good clinical net benefit of the two nomograms in both training and validation cohorts. Kaplan-Meier survival analyses showed superior discrimination of DFS and OS in the new risk stratification system compared with the AJCC stage system. Finally, postoperative patients with PGC who underwent adjuvant radiotherapy had a better prognosis in the high-, and medium-risk subgroups (p < 0.05), but not for the low-risk subgroup.ConclusionThe immune-inflammatory-nutrition indicators and ACCI played an important role in both DFS and OS of PGC patients. Adjuvant radiotherapy had no benefit in the low-risk subgroup for PGC patients who underwent surgical resection. The newly established nomogram models perform well and can provide an individualized prognostic reference, which may be helpful for patients and surgeons in proper follow-up strategies.

Prediction of Adolescents' Fluid Intelligence Scores based on Deep Learning with Reconstruction Regularization.

The aim of this study was to develop a predictive model for uncorrected/actual fluid intelligence scores in 9-10 year old children using magnetic resonance T1-weighted imaging. Explore the predictive performance of an autoencoder model based on reconstruction regularization for fluid intelligence in adolescents. We collected actual fluid intelligence scores and T1-weighted MRIs of 11,534 adolescents who completed baseline tasks from ABCD Data Release 3.0. A total of 148 ROIs were selected and 604 features were proposed by FreeSurfer segmentation. The training and testing sets were divided in a ratio of 7:3. To predict fluid intelligence scores, we used AE, MLP and classic machine learning models, and compared their performance on the test set. In addition, we explored their performance across gender subpopulations. Moreover, we evaluated the importance of features using the SHapley Additive Explain method. Results: The proposed model achieves optimal performance on the test set for predicting actual fluid intelligence scores (PCC = 0.209 ± 0.02, MSE = 105.212 ± 2.53). Results show that autoencoders with refactoring regularization are significantly more effective than MLPs and classical machine learning models. In addition, all models performed better on female adolescents than on male adolescents. Further analysis of relevant characteristics in different populations revealed that this may be related to gender differences in underlying fluid intelligence mechanisms. We construct a weak but stable correlation between brain structural features and raw fluid intelligence using autoencoders. Future research may need to explore ensemble regression strategies utilizing multiple machine learning algorithms on multimodal data in order to improve the predictive performance of fluid intelligence based on neuroimaging features.

Assessing Machine Learning and Statistical Methods for Rock Glacier‐Based Permafrost Distribution in Northern Kargil Region

ABSTRACTEstimating permafrost distribution in high‐mountain areas is challenging. In these situations, rock glaciers, provide valuable insights into permafrost distribution and are often used as proxies for identifying permafrost occurrence. Integrating various climatological and topographical conditioning factors with rock glaciers enables inferring the distribution of permafrost in these environments. This study utilized three machine learning models such as random forest (RF), support vector (SVM), and artificial neural network (ANN), and one statistical model, namely, the frequency ratio (FR), to assess the permafrost probability over the northern Kargil region of Indian Himalayas. Among 198 rock glaciers identified through high‐resolution images from Google Earth, 70% are used as training dataset, rest 30% as testing dataset. The study considered eight factors: slope, aspect, elevation, curvature, mean annual land surface temperature (MA‐LST), mean annual normalized difference snow index (MA‐NDSI), mean annual normalized difference water index (MA‐NDWI), and lithology for mapping. Furthermore, the SHapley Additive exPlanations (SHAP) test assessed the variable importance for model performance. The results revealed that the RF model performs best for permafrost probability mapping, followed by the SVM, FR, and ANN models. The study also found that 11% of the total geographic area has a high and very high probability of permafrost occurrence.

A hybrid CNN–RNN model for rainfall–runoff modeling in the Potteruvagu watershed of India

AbstractAccurate rainfall‐runoff analysis is essential for water resource management, with artificial intelligence (AI) increasingly used in this and other hydrological areas. The need for precise modelling has driven substantial advancements in recent decades. This study employed six AI models. These were the support vector regression model (SVR), the multilinear regression model (MLR), the extreme gradient boosting model (XGBoost), the long‐short‐term memory (LSTM) model, the convolutional neural network (CNN) model, and the convolutional recurrent neural network (CNN‐RNN) hybrid model. It covered 1998–2006, with 1998–2004 for calibration/training and 2005–2006 for validation/testing. Five metrics were used to measure model performance: coefficient of determination (R2), Nash‐Sutcliffe efficiency (NSE), mean absolute error (MAE), root‐mean square error (RMSE), and RMSE‐observations standard deviation ratio (RSR). The hybrid CNN‐RNN model performed best in both training and testing periods (training: R2 is 0.92, NSE is 0.91, MAE is 10.37 m3s−1, RMSE is 13.13 m3s−1, and RSR is 0.30; testing: R2 is 0.95, NSE is 0.94, MAE is 12.18 m3s−1, RMSE is 15.86 m3s−1, and RSR is 0.25). These results suggest the hybrid CNN‐RNN model is highly effective for rainfall‐runoff analysis in the Potteruvagu watershed.

Spatiotemporal Modeling of Carbon Fluxes over Complex Underlying Surfaces along the North Shore of Hangzhou Bay

Urban areas contribute to over 80% of carbon dioxide emissions, and considerable efforts are being undertaken to characterize spatiotemporal variations of CO2 (carbon dioxide) at a city, regional, and national level, aiming at providing pipelines for carbon mission reduction. The complex underlying surface composition of urban areas makes process-based and physiology-based models inadequate for simulating carbon flux in this context. In this study, long short-term memory (LSTM), support vector machine (SVM), random forest (RF), and artificial neural network (ANN) were employed to develop and investigate their viability in estimating carbon flux at the ecosystem level. All the data used in our study were derived from the long-term chronosequence observations collected from the flux towers within urban complex underlying surface, along with meteorological reanalysis datasets. To assess the generalization ability of these models, the following statistical metrics were utilized: coefficient of determination (R2), root mean square error (RMSE), and mean absolute error (MAE). Our analysis revealed that the RF model performed the best in simulating carbon flux over long time series, with the highest R2 values reaching up to 0.852, and exhibiting the smallest RMSE and MAE values at 0.293 μmol·m−2·s−1 and 0.157 μmol·m−2·s−1. As a result, the RF model was chosen for simulating carbon flux at spatial scale and assessing the impact of urban impervious surfaces in the simulation. The results showed that the RF model performs well in simulating carbon flux at the spatial scale. The input of impervious surface area index can improve the performance of the RF model in simulating carbon flux, with R2 values of 84.46% (with the impervious surface area index in) and 83.74% (without the impervious surface area index in). Furthermore, the carbon flux in Fengxian District, Shanghai, exhibited significant spatial heterogeneity: the CO2 flux in the western part of Fengxian District was less than in the eastern part, and the CO2 flux gradually increased from the west to the east. In addition, we creatively introduced the diurnal impervious surface area index based on the Kljun model, and clarified the influence of impervious surface on the spatiotemporal simulation of CO2 flux over the complex urban underlying surface. Based on these findings, we conclude that the RF models can be effectively applied for estimating carbon flux on the complex underlying urban surface. The results of our study reduce the uncertainty in modeling carbon cycling in terrestrial ecosystems, and make the variety of models for the carbon cycling of terrestrial ecosystems more diverse.

Assessment of two approaches for very short range precipitation prediction for a convection-dominant period at different scales

Extreme precipitation plays a crucial role in managing weather-related risks, particularly flooding, and has a significant impact on various sectors such as hydrology, agriculture, and economy. It is important to evaluate precipitation forecast outputs to help refining prediction models, especially for heavy events. This study examines the performance of two weather models, both designed to provide frequently updated forecasts for a short time range, but using different approaches: extrapolation-based and fully dynamical 3D numerical weather prediction (NWP) models. These models are assessed over a two-month period with several heavy thunderstorms, using data from Austria and a smaller catchment in southeast Austria. We addressed the errors and uncertainties in these models using different verification metrics and two different observational datasets. In general, both models performed well in the first hour of the forecast at both large and small scales. However, performance decreased with lead time after the first hour, with a faster decline in the extrapolation-based model. The performance of the models changed when considering different precipitation thresholds. For higher thresholds (heavier precipitation), the NWP model performed better than the extrapolation-based model, whereas it performed worse for lower thresholds. When using the object-oriented Structure Amplitude Location (SAL) metric, the dynamical NWP model performed better regarding the structure of precipitation, while the extrapolated-based model had a better performance when considering the amplitude parameter, i.e., the total amount of precipitation. With a lower error in the first hours of each run, the dynamical NWP model showed that an hourly update cycle helps to reduce the forecast errors noticeably. The results highlight the need of considering forecast model performance at different intensity thresholds and spatial scales.

Developing and comparing deep learning and machine learning algorithms for osteoporosis risk prediction.

Osteoporosis, characterized by low bone mineral density (BMD), is an increasingly serious public health issue. So far, several traditional regression models and machine learning (ML) algorithms have been proposed for predicting osteoporosis risk. However, these models have shown relatively low accuracy in clinical implementation. Recently proposed deep learning (DL) approaches, such as deep neural network (DNN), which can discover knowledge from complex hidden interactions, offer a new opportunity to improve predictive performance. In this study, we aimed to assess whether DNN can achieve a better performance in osteoporosis risk prediction. By utilizing hip BMD and extensive demographic and routine clinical data of 8,134 subjects with age more than 40 from the Louisiana Osteoporosis Study (LOS), we developed and constructed a novel DNN framework for predicting osteoporosis risk and compared its performance in osteoporosis risk prediction with four conventional ML models, namely random forest (RF), artificial neural network (ANN), k-nearest neighbor (KNN), and support vector machine (SVM), as well as a traditional regression model termed osteoporosis self-assessment tool (OST). Model performance was assessed by area under 'receiver operating curve' (AUC) and accuracy. By using 16 discriminative variables, we observed that the DNN approach achieved the best predictive performance (AUC = 0.848) in classifying osteoporosis (hip BMD T-score ≤ -1.0) and non-osteoporosis risk (hip BMD T-score > -1.0) subjects, compared to the other approaches. Feature importance analysis showed that the top 10 most important variables identified by the DNN model were weight, age, gender, grip strength, height, beer drinking, diastolic pressure, alcohol drinking, smoke years, and economic level. Furthermore, we performed subsampling analysis to assess the effects of varying number of sample size and variables on the predictive performance of these tested models. Notably, we observed that the DNN model performed equally well (AUC = 0.846) even by utilizing only the top 10 most important variables for osteoporosis risk prediction. Meanwhile, the DNN model can still achieve a high predictive performance (AUC = 0.826) when sample size was reduced to 50% of the original dataset. In conclusion, we developed a novel DNN model which was considered to be an effective algorithm for early diagnosis and intervention of osteoporosis in the aging population.

Development of a deep learning model for predicting recurrence of hepatocellular carcinoma after liver transplantation.

Liver transplantation (LT) is one of the main curative treatments for hepatocellular carcinoma (HCC). Milan criteria has long been applied to candidate LT patients with HCC. However, the application of Milan criteria failed to precisely predict patients at risk of recurrence. As a result, we aimed to establish and validate a deep learning model comparing with Milan criteria and better guide post-LT treatment. A total of 356 HCC patients who received LT with complete follow-up data were evaluated. The entire cohort was randomly divided into training set (n = 286) and validation set (n = 70). Multi-layer-perceptron model provided by pycox library was first used to construct the recurrence prediction model. Then tabular neural network (TabNet) that combines elements of deep learning and tabular data processing techniques was utilized to compare with Milan criteria and verify the performance of the model we proposed. Patients with larger tumor size over 7 cm, poorer differentiation of tumor grade and multiple tumor numbers were first classified as high risk of recurrence. We trained a classification model with TabNet and our proposed model performed better than the Milan criteria in terms of accuracy (0.95 vs. 0.86, p < 0.05). In addition, our model showed better performance results with improved AUC, NRI and hazard ratio, proving the robustness of the model. A prognostic model had been proposed based on the use of TabNet on various parameters from HCC patients. The model performed well in post-LT recurrence prediction and the identification of high-risk subgroups.

Evaluation of CERES –Rice Model for Simulating Rice Yield and Phenophases

In this study, the performance of the CERES-Rice model in simulating the growth and yield of the Rajeshwari variety in the Raipur district of Chhattisgarh, India, was evaluated. Utilizing observed data from 2021 and 2022, the model was calibrated and validated using key parameters such as days to anthesis, physiological maturity, and yield. Calibration involved adjusting genetic coefficients to improve simulation accuracy, with validation ensuring the model's reliability beyond the calibration period. The comparison between observed and simulated data for crop performance parameters showed that the model performed reasonably well. For days to Anthesis, the RMSE was 4.32 with a d-stat of 0.59, and an error of 5.4%. For Days to Panicle initiation, the RMSE was 1.83, the d-stat was 0.82, and the error was -4.7%. For days to PM, the RMSE was 6.7, the d-stat was 0.65, and the error was 3.0%. Yield showed an RMSE of 472.4, a d-stat of 0.81, and an error of 7.7%. F The mean simulated values closely matched the observed means, indicating overall good model accuracy. In this study, fine tuning the genetic coefficients of CERES rice model for the variety Rjeshwari was done and can be used for further studies and applications.

Global hydrological models continue to overestimate river discharge

Global hydrological models (GHMs) are widely used to assess the impact of climate change on streamflow, floods, and hydrological droughts. For the ‘model evaluation and impact attribution’ part of the current round of the Inter-Sectoral Impact Model Intercomparison Project (ISIMIP3a), modelling teams generated historical simulations based on observed climate and direct human forcings with updated model versions. Here we provide a comprehensive evaluation of daily and maximum annual discharge based on ISIMIP3a simulations from nine GHMs by comparing the simulations to observational data from 644 river gauge stations. We also assess low flows and the effects of different river routing schemes. We find that models can reproduce variability in daily and maximum annual discharge, but tend to overestimate both quantities, as well as low flows. Models perform better at stations in wetter areas and at lower elevations. Discharge routed with the river routing model CaMa-Flood can improve the performance of some models, but for others, variability is overestimated, leading to reduced model performance. This study indicates that areas for future model development include improving the simulation of processes in arid regions and cold dynamics at high elevations. We further suggest that studies attributing observed changes in discharge to historical climate change using the current model ensemble will be most meaningful in humid areas, at low elevations, and in places with a regular seasonal discharge as these are the regions where the underlying dynamics seem to be best represented.

Modified cotton gauze with high hemostatic efficacy due to controllable hygroscopicity and wet tissue adhesiveness

AbstractIn circumstances of severe traumatic bleeding, urgent bleeding control is vital to the wellness and life of the wounded person. Among many hemostatic measures, gauze is widely used because of cost‐effectiveness, safety and efficiency. However, it may suck a large amount of blood before bleeding stops. Thus, advanced hemostatic gauzes less blood loss and quick hemostatic rate and is greatly demanded. Herein, a novel tissue adhesive gauze (TA‐gauze) is developed by coating a thin layer of polyethyleneimine/poly(acrylic acid) (PEI/PAA) on the plain cotton gauze. This TA‐gauze has a reduced hydrophilicity and gets a new trait of wet tissue adhesiveness from the PEI/PAA coating layer. The TA‐gauze shows much improved hemostatic performance in the rat injury models, that is, its blood loss in the rat femoral artery model is just 13% that of the cotton gauze, due to it can retard the blood movement in the gauze and at the seam of gauze/tissue through a synergistic effect of tissue adhesion and hygroscopicity. It is biocompatible and on‐demand detachable. The simple preparation procedure and high hemostatic efficiency of the TA‐gauze may promise a bright future for an ideal hemostatic dressing for bleeding control.

Predicting Healthy Start Scheme Uptake using Deprivation and Food Insecurity Measures.

Introduction &amp; BackgroundThe level of food insecurity in England is widening, with low-income families requiring more support to reduce income inequalities. The government have introduced policies to address these issues with targeted subsidies on healthy food on programs such as the Healthy Start Scheme. Despite this, national uptake of the Healthy Start Scheme remains lower than the government target. Objectives &amp; ApproachOur study aims to predict uptake and take up discrepancies at a local authority level and understand the measures contributing to the prediction using anonymised supermarket loyalty card data records for over 4 million customers, deprivation and food insecurity measures. We used a machine-learning approach utilising transactional data, ONS Index of Deprivation datasets, neighbourhood statistics, and NHS Healthy Start Scheme uptake data. Regression prediction models were used to evaluate and predict the outcomes, whilst feature importance tools were used to evaluate the variables weighing within the model. Relevance to Digital FootprintsThis study leverages transaction data from a UK retailer to understand lifestyle factors at a local authority level and assesses their usefulness in predicting the scheme’s uptake. Loyalty card transactional data can provide valuable insight into purchase behaviour linked to health and nutrition. ResultsThe Linear and Ridge Regression models performed better than other prediction models. Analysis of measures revealed that whilst deprivation and population-related measures had a high contribution to the prediction model, findings from transactional data measures provided valuable insight into shopping behavioural characteristics that contribute to the model performance. Results suggested that areas with higher spending on fruits and vegetables and high-calorie food were associated with higher uptake prediction in test data but the converse for high spend on fish. Conclusions &amp; ImplicationsOur study indicates that shopping data measures such as spend on fruits and vegetables, high-calorie food, fish and products bought can be utilised for prediction models for uptake and take-up discrepancy of the Healthy Start Scheme. This study highlights the complexity of understanding factors influencing public policy effectiveness and the need for tailored approaches in diverse urban contexts.

Can Pretreatment MRI and Planning CT Radiomics Improve Prediction of Complete Pathological Response in Locally Advanced Rectal Cancer Following Neoadjuvant Treatment?

The treatment response to neoadjuvant chemoradiation (nCRT) differs largely in individuals treated for rectal cancer. In this study, we investigated the role of radiomics to predict the pathological response in locally advanced rectal cancers at different treatment time points: (1) before thestart of any treatment using baseline T2-weighted MRI (T2W-MR) and (2) at the start of radiation treatment using planning CT. Patients on nCRT followed by surgery between June 2017 to December 2019 were included in the study. Histopathological tumour response grading (TRG) was used for classification, and gross tumour volume was defined by the radiation oncologists. Following resampling, 100 and 103 pyradiomic features were extracted from T2W-MR and planning CT images, respectively. Synthetic minority oversampling technique (SMOTE) was used to address class imbalance. Four machine learning classifiers built clinical, radiomic, and merged models. Model performances were evaluated on a held-out test dataset following 3-fold cross-validation using area under the receiver operator characteristic curves (AUC) with bootstrap 95% confidence intervals. One hundred and fiftypatients were included; 58/150 with TRG 1 were classified as complete responders, and rest were incomplete responders (IR). Clinical models performed better (AUC = 0.68) compared to radiomics models (AUC = 0.62). Overall, the clinical + T2W-MR model showed best performance (AUC = 0.72) in predicting the pathological response prior to therapy. Clinical + Planning CT-merged models could only achieve the highest AUC of 0.66. Merging clinical and baseline T2W-MR radiomics enhances predicting pathological response in rectal cancer. Validation in larger cohorts is warranted, especially for watch and wait strategies.

Performance analysis of Bayesian optimised gradient-boosted decision trees for digital elevation model (DEM) error correction: interim results

Abstract. Gradient-Boosted Decision Trees (GBDTs), particularly when tuned with Bayesian optimisation, are powerful machine learning techniques known for their effectiveness in handling complex, non-linear data. However, the performance of these models can be significantly influenced by the characteristics of the terrain being analysed. In this study, we assess the performance of three Bayesian-optimised GBDTs (XGBoost, LightGBM and CatBoost) using digital elevation model (DEM) error correction as a case study. The performance of the models is investigated across five landscapes in Cape Town South Africa: urban/industrial, agricultural, mountain, peninsula and grassland/shrubland. The models were trained using a selection of datasets (elevation, terrain parameters and land cover). The comparison entailed an analysis of the model execution times, regression error metrics, and level of improvement in the corrected DEMs. Generally, the optimised models performed considerably well and demonstrated excellent predictive capability. CatBoost emerged with the best results in the level of improvement recorded in the corrected DEMs, while LightGBM was the fastest of all models in the execution time for Bayesian optimisation and model training. These findings offer valuable insights for applying machine learning and hyperparameter tuning in remote sensing.

Advancing Fuel Spray Characterization: A Machine Learning Approach for Directly Injected Gasoline Fuel Sprays

Compression ignition engines when operated on gasoline fuels cause significant reduction in NOx and particulate emissions. In such advanced combustion strategy, the fuel-oxidiser mixing process, intensified by the prolonged ignition delays of gasoline fuels, directly affects the stability and efficiency of combustion. Thus, optimising fuel spray characteristics leads to optimisation of injector design, engine performance and subsequent decrease in emissions. Since spray development is a complex process that involves wide range of length and time scales, computationally expensive modelling techniques can be replaced with machine learning (ML) models. These ML models are employed to predict spray characteristics utilizing datasets generated from comprehensive spray studies. In this work, a dataset of about 5400 instances taken from non-evaporating gasoline fuel spray imaging experiments under Gasoline Compression Injection (GCI) engine conditions is used to train various ML models with data split of 70 % and 30 % for training and testing, respectively, with five-folds cross-validation performed within the training. The fuel injection pressure (60 – 150 MPa), chamber pressure (0.1 – 2 MPa), nozzle diameter, nozzle hole conicity and injection duration are used as input features to the models for predicting the spray tip penetration and spray angle. The performance of four ML models was evaluated and compared under default and tuned hyperparameters against experimental data and available physics-based correlations in the literature. The models include random forest, extreme gradient boosting, multilayer perceptron, and elastic-net. The results show that the hyperparameter-tuned extreme gradient boosting model performs best in predicting the spray parameters. The overall model performance was evaluated using the coefficient of determination (R2), mean absolute error (MAE), and root mean squared error (RMSE), resulting in values of 0.884, 0.651, and 1.571, respectively. This study presents compelling evidence demonstrating the effectiveness of ML as a powerful tool for isolating non-linear behaviors from physical processes. By effectively decoupling these behaviors, ML enhances the accuracy of predicting spray characteristics while significantly reducing computational costs. The application of ML in fuel injector design has the potential to revolutionize engine performance and contribute to substantial reductions in emissions.

Comparison of Fractional-Order Monkeypox Model with Singular and Non-Singular Kernels

The singularity of the kernel of the Caputo fractional derivative has become an issue, leading many researchers to consider the Atangana-Baleanu-Caputo (ABC) fractional derivative in epidemic models where the kernel is non-singular. In this context, the objective of this study is to compare the calibration and forecasting performance of fractional-order monkeypox models with singular and nonsingular kernels, represented by the model with respect to the Caputo operator and the ABC operator, respectively. We have proposed a monkeypox epidemic model with respect to the ABC operator (MPXABC), where the model with respect to the Caputo derivative (MPXC) has been proposed in previous research. We have analyzed the existence and uniqueness of the solution. Three equilibrium points of the model are endemic, human endemic, and monkeypox-free, and their global stability has been investigated. The global dynamics of the MPXABC are the same as those of the MPXC. In evaluating the performance, we collected secondary data on weekly monkeypox cases from June 1 to November 23, 2022, in the USA. Parameter estimation has been performed using the least squares method, while the solutions of the model have been determined numerically using a predictor-corrector scheme. The benchmark for performance has been determined based on the root mean square error. Data calibration and forecasting indicate that the MPXC generally has the best performance for each value of the derivative order. For certain values of derivative order, the MPXABC performs better than the corresponding firstorder model. However, generally, the corresponding first-order model performs better than the MPXABC. Depending on the data trends and the specified orders, the MPXC outperforms the MPXABC. Thus, the singularity issue of the Caputo derivative does not always have a negative impact on model fitting to data.

Machine Learning-Based Prediction of Helicobacter pylori Infection Study in Adults.

BACKGROUND Helicobacter pylori has a high infection rate worldwide, and epidemiological study of H. pylori is important. Artificial intelligence has been widely used in the field of medical research and has become a hotspot in recent years. This paper proposed a prediction model for H. pylori infection based on machine learning in adults. MATERIAL AND METHODS Adult patients were selected as research participants, and information on 30 factors was collected. The chi-square test, mutual information, ReliefF, and information gain were used to screen the feature factors and establish 2 subsets. We constructed an H. pylori infection prediction model based on XGBoost and optimized the model using a grid search by analyzing the correlation between features. The performance of the model was assessed by comparing its accuracy, recall, precision, F1 score, and AUC with those of 4 other classical machine learning methods. RESULTS The model performed better on the part B subset than on the part A subset. Compared with the other 4 machine learning methods, the model had the highest accuracy, recall, F1 score, and AUC. SHAP was used to evaluate the importance of features in the model. It was found that H. pylori infection of family members, living in rural areas, poor washing hands before meals and after using the toilet were risk factors for H. pylori infection. CONCLUSIONS The model proposed in this paper is superior to other models in predicting H. pylori infection and can provide a scientific basis for identifying the population susceptible to H. pylori and preventing H. pylori infection.

Automated control of Impella maintains optimal left ventricular unloading during periods of unstable hemodynamics and prevents myocardial damage in acute myocardial infarction

BackgroundLeft ventricular (LV) unloading by Impella, an intravascular microaxial pump, has been shown to exert dramatic cardioprotective effects in acute clinical settings of cardiovascular diseases. Total Impella support (no native LV ejection) is far more efficient in reducing LV energetic demand than partial Impella support, but the manual control of pump speed to maintain stable LV unloading is difficult and impractical. We aimed to develop an Automatic IMpella Optimal Unloading System (AIMOUS), which controls Impella pump speed to maintain LV unloading degree using closed-feedback control. We validated the AIMOUS performance in an animal model. MethodsIn dogs, we identified the transfer function from pump speed to LV systolic pressure (LVSP) under total support conditions (n = 5). Using the transfer function, we designed the feedback controller of AIMOUS to keep LVSP at 40 mmHg and examined its performance by volume perturbations (n = 9). Lastly, AIMOUS was applied in the acute phase of ischemia-reperfusion in dogs. Four weeks after ischemia-reperfusion, we assessed LV function and infarct size (n = 10). ResultsAIMOUS maintained constant LVSP, thereby ensuring a stable LV unloading condition regardless of volume withdrawal or infusion (±8 ml/kg from baseline). AIMOUS in the acute phase of ischemia-reperfusion markedly improved LV function and reduced infarct size (No Impella support: 13.9 ± 1.3 vs. AIMOUS: 5.7 ± 1.9%, P < 0.05). ConclusionsAIMOUS is capable of maintaining optimal LV unloading during periods of unstable hemodynamics. Automated control of Impella pump speed in the acute phase of ischemia-reperfusion significantly reduced infarct size and prevented subsequent worsening of LV function.

Exploratory assessment of the effect of systemic administration of soluble glycoprotein 130 on cognitive performance and chemokine levels in a mouse model of experimental traumatic brain injury

Uncontrolled neuroinflammation mediates traumatic brain injury (TBI) pathology and impairs recovery. Interleukin-6 (IL-6), a pleiotropic inflammatory regulator, is associated with poor clinical TBI outcomes. IL-6 operates via classical-signaling through membrane-bound IL-6 receptor (IL-6R) and trans-signaling through soluble IL-6 receptor (s)IL-6R. IL-6 trans-signaling specifically contributes to neuropathology, making it a potential precision therapeutic TBI target. Soluble glycoprotein 130 (sgp130) prevents IL-6 trans-signaling, sparing classical signaling, thus is a possible treatment. Mice received either controlled cortical impact (CCI) (6.0 ± 0.2 m/s; 2 mm; 50-60ms) or sham procedures. Vehicle (VEH) or sgp130-Fc was subcutaneously administered to sham (VEH or 1 µg) and CCI (VEH, 0.25 µg or 1 µg) mice on days 1, 4, 7, 10 and 13 post-surgery to assess effects on cognition [Morris Water Maze (MWM)] and ipsilateral hemisphere IL-6 related biomarkers (day 21 post-surgery). CCI + sgp130-Fc groups (0.25 µg and 1 µg) were combined for analysis given similar behavior/biomarker outcomes. CCI + VEH mice had longer latencies and path lengths to the platform and increased peripheral zone time versus Sham + VEH and Sham + sgp130-Fc mice, suggesting injury-induced impairments in learning and anxiety. CCI + sgp130-Fc mice had shorter platform latencies and path lengths and had decreased peripheral zone time, indicating a therapeutic benefit of sgp130-Fc after injury on learning and anxiety. Interestingly, Sham + sgp130-Fc mice had shorter platform latencies, path lengths and peripheral zone times than Sham + VEH mice, suggesting a beneficial effect of sgp130-Fc, independent of injury. CCI + VEH mice had increased brain IL-6 and decreased sgp130 levels versus Sham + VEH and Sham + sgp130-Fc mice. There was no treatment effect on IL-6, sIL6-R or sgp130 in Sham + VEH versus Sham + sgp130-Fc mice. There was also no treatment effect on IL-6 in CCI + VEH versus CCI + sgp130-Fc mice. However, CCI + sgp130-Fc mice had increased sIL-6R and sgp130 versus CCI + VEH mice, demonstrating sgp130-Fc treatment effects on brain biomarkers. Inflammatory chemokines (MIP-1β, IP-10, MIG) were increased in CCI + VEH mice versus Sham + VEH and Sham + sgp130-Fc mice. However, CCI + sgp130-Fc mice had decreased chemokine levels versus CCI + VEH mice. IL-6 positively correlated, while sgp130 negatively correlated, with chemokine levels. Overall, we found that systemic sgp130-Fc treatment after CCI improved learning, decreased anxiety and reduced CCI-induced brain chemokines. Future studies will explore sex-specific dosing and treatment mechanisms for sgp130-Fc therapy.