The uneven distribution of the temperature field in the track structure,caused by various meteorological factors such as extremely low temperatures and snowfall,leads to significant temperature loads and is the primar...The uneven distribution of the temperature field in the track structure,caused by various meteorological factors such as extremely low temperatures and snowfall,leads to significant temperature loads and is the primary cause of damage to China Railway Track System(CRTS)III ballastless tracks in cold regions during service.In this study,to predict the temperature of the track structure accurately,we analyzed meteorological data collected from Shenyang,China,and identified the factors that had the most effect on the track temperature field.We propose a temporal convolutional network(TCN)-based temperature field prediction model for ballastless tracks(TCN-Track model),which enhances the ability to extract and fuse local and global features from complex long-term meteorological data.The results indicate that the proposed TCN-Track model performs well in predicting track temperature fields from meteorological data,with a mean absolute error(MAE)ranging from 0.26 to 0.39,a root mean square error(RMSE)ranging from 0.32 to 0.50,and correlation coefficient(R)values ranging from 0.888 to 0.985.Compared with a long short-term memory(LSTM)model,the MAE of the TCN-Track model is reduced by 89.17%and the RMSE by 88.51%.This method offers a new solution for accurately predicting the temperature field of ballastless tracks in cold regions,aiding in predicting and preventing track damage caused by low temperatures.展开更多
Northeast China serves as an important crop production region.Accurately forecasting summer precipitation in Northeast China(NEC-PR)has been a challenge due to its wide range of time scales influenced by varying clima...Northeast China serves as an important crop production region.Accurately forecasting summer precipitation in Northeast China(NEC-PR)has been a challenge due to its wide range of time scales influenced by varying climatic conditions.This study presents a scale separation hybrid statistical model with recurrent neural network(SS-RNN)to predict the summer monthly NEC-PR.The SS-RNN model decomposes the multiple scales of the NEC-PR into several spatiotemporal intrinsic mode functions covering annual to decadal time scales.This strategy provides a way to derive appropriate predictors and establish predictive models for the primary spatial modes of the NEC-PR at various time scales.Our results demonstrate substantial improvements by the SS-RNN model in predicting the summer monthly NEC-PR as compared with dynamic models,particularly in predicting the spatial pattern of the NEC-PR.In this paper we take August,the month of the highest NEC-PR,to assess our model skill.Independent forecasts of the August NEC-PR over the period 2021–24 achieve significant spatial anomaly correlation coefficients,reaching a maximum value of 0.83.Additional verifications by station observations show that the model hits most station anomalies,achieving a mean predictive skill score of 90.展开更多
To investigate the influence of coarse aggregate parent rock properties on the elastic modulus of concrete,the mineralogical properties and stress-strain curves of granite and dolomite parent rocks,as well as the stre...To investigate the influence of coarse aggregate parent rock properties on the elastic modulus of concrete,the mineralogical properties and stress-strain curves of granite and dolomite parent rocks,as well as the strength and elastic modulus of mortar and concrete prepared with mechanism aggregates of the corresponding lithology,and the stress-strain curves of concrete were investigated.In this paper,a coarse aggregate and mortar matrix bonding assumption is proposed,and a prediction model for the elastic modulus of mortar is established by considering the lithology of the mechanism sand and the slurry components.An equivalent coarse aggregate elastic modulus model was established by considering factors such as coarse aggregate particle size,volume fraction,and mortar thickness between coarse aggregates.Based on the elastic modulus of the equivalent coarse aggregate and the remaining mortar,a prediction model for the elastic modulus of the two and three components of concrete in series and then in parallel was established,and the predicted values differed from the measured values within 10%.It is proposed that the coarse aggregate elastic modulus in highstrength concrete is the most critical factor affecting the elastic modulus of concrete,and as the coarse aggregate elastic modulus increases by 27.7%,the concrete elastic modulus increases by 19.5%.展开更多
Customer churn is the rate at which customers discontinue doing business with a company over a given time period.It is an essential measure for businesses to monitor high churn rates,as they often indicate underlying ...Customer churn is the rate at which customers discontinue doing business with a company over a given time period.It is an essential measure for businesses to monitor high churn rates,as they often indicate underlying issues with services,products,or customer experience,resulting in considerable income loss.Prediction of customer churn is a crucial task aimed at retaining customers and maintaining revenue growth.Traditional machine learning(ML)models often struggle to capture complex temporal dependencies in client behavior data.To address this,an optimized deep learning(DL)approach using a Regularized Bidirectional Long Short-Term Memory(RBiLSTM)model is proposed to mitigate overfitting and improve generalization error.The model integrates dropout,L2-regularization,and early stopping to enhance predictive accuracy while preventing over-reliance on specific patterns.Moreover,this study investigates the effect of optimization techniques on boosting the training efficiency of the developed model.Experimental results on a recent public customer churn dataset demonstrate that the trained model outperforms the traditional ML models and some other DL models,such as Long Short-Term Memory(LSTM)and Deep Neural Network(DNN),in churn prediction performance and stability.The proposed approach achieves 96.1%accuracy,compared with LSTM and DNN,which attain 94.5%and 94.1%accuracy,respectively.These results confirm that the proposed approach can be used as a valuable tool for businesses to identify at-risk consumers proactively and implement targeted retention strategies.展开更多
Pinus radiata(D.Don)dominates New Zealand's forestry industry,constituting 91%of plantations,and is among the world's most important plantation species.Given the socio-economic and environmental importance of ...Pinus radiata(D.Don)dominates New Zealand's forestry industry,constituting 91%of plantations,and is among the world's most important plantation species.Given the socio-economic and environmental importance of this species,it is important to have accurate and precise projections over time to make efficient decisions for forest management and greenfield investments in afforestation projects,especially for permanent carbon forests.Future projections of any natural resource systems rely on modeling;however,the acceleration of climate change makes future projections of yield less certain.These challenges also impact national expectations of the contribution planted forests will provide to address climate change and meet international commitments under the Paris Agreement.Using a large national-scale set of contemporary ground-measured data(2013–2023),this study investigates the performance of two growth models developed over 30 years ago that are widely used by NZ plantation growers:1)the Pumice Plateau Model 1988(PPM88)and 2)the 300-index(including a model variant of regional drift).Model simulations were made using the FORECASTER modeling suite with geographic boundaries to adjust for drift in space and time.Basal area(BA,m^(2)⋅ha^(-1))and volume(m^(3)⋅ha^(-1))were simulated,and standard errors and goodness-of-fit metrics calculated up to a typical rotation age of 30 years.Model residuals were then separated and analysed for the main plantation growing regions.The models overpredicted observed growth by between 6.8%and 16.2%,but model predictions and errors varied significantly between regions.The results of this study provided clear evidence of divergence between the outputs of both models and the measured data.Finally,this study suggests future measures to address challenges posed by these discrepancies that will provide better information for forest management and investment decisions in a changing climate.展开更多
Objective:To investigate the impact of targeted nursing interventions based on frailty prediction models on peri-hospitalization clinical outcomes in middle-aged and elderly patients with upper gastrointestinal bleedi...Objective:To investigate the impact of targeted nursing interventions based on frailty prediction models on peri-hospitalization clinical outcomes in middle-aged and elderly patients with upper gastrointestinal bleeding(UGIB).Methods:A prospective cohort study was conducted,and 126 middle-aged and elderly patients with UGIB admitted from August 2024 to August 2025 were selected as the study subjects.The patients were divided into the intervention group(63 cases)and the control group(63 cases)based on whether they received nursing intervention based on frailty prediction models.The control group received routine care,while the intervention group,on the basis of routine care,used the FRAIL scale combined with laboratory indicators(albumin,hemoglobin,etc.)to establish a predictive model to evaluate patients within 24 hours of admission,and implemented multi-dimensional targeted nursing intervention for pre-frailty or frailty patients screened out.The incidence of frailty,rebleeding rate,average length of stay,hospitalization cost,and nursing satisfaction during hospitalization were compared between the two groups.Results:The incidence of frailty during hospitalization in the intervention group was 11.1%(7 cases/63 cases),significantly lower than 31.7%(20 cases/63 cases)in the control group,and the difference was statistically significant(p<0.05).The rebleeding rate of 4.8%vs 12.7%,the average length of stay of(7.2±1.5)days vs(9.1±2.2)days,and the average hospitalization cost of(23,000±6,000)yuan vs(28,000±7,000)yuan in the intervention group were all lower than those in the control group(all p<0.05).The nursing satisfaction score of the intervention group(93.5±4.2)points was higher than that of the control group(86.3±5.8)points(p<0.05).Conclusion:The frailty prediction model applied to the peri-hospitalization care of middle-aged and elderly patients with UGIB can effectively identify frailty risk.Through early targeted intervention,the incidence of frailty and rebleeding rate can be reduced,the length of hospital stay can be shortened,medical expenses can be reduced,and nursing satisfaction can be improved,which has clinical promotion value.展开更多
Machine learning-assisted methods for rapid and accurate prediction of temperature field,mushy zone,and grain size were proposed for the heating−cooling combined mold(HCCM)horizontal continuous casting of C70250 alloy...Machine learning-assisted methods for rapid and accurate prediction of temperature field,mushy zone,and grain size were proposed for the heating−cooling combined mold(HCCM)horizontal continuous casting of C70250 alloy plates.First,finite element simulations of casting processes were carried out with various parameters to build a dataset.Subsequently,different machine learning algorithms were employed to achieve high precision in predicting temperature fields,mushy zone locations,mushy zone inclination angle,and billet grain size.Finally,the process parameters were quickly optimized using a strategy consisting of random generation,prediction,and screening,allowing the mushy zone to be controlled to the desired target.The optimized parameters are 1234℃for heating mold temperature,47 mm/min for casting speed,and 10 L/min for cooling water flow rate.The optimized mushy zone is located in the middle of the second heat insulation section and has an inclination angle of roughly 7°.展开更多
BACKGROUND:This study aims to develop and validate a machine learning-based in-hospital mortality predictive model for acute aortic syndrome(AAS)in the emergency department(ED)and to derive a simplifi ed version suita...BACKGROUND:This study aims to develop and validate a machine learning-based in-hospital mortality predictive model for acute aortic syndrome(AAS)in the emergency department(ED)and to derive a simplifi ed version suitable for rapid clinical application.METHODS:In this multi-center retrospective cohort study,AAS patient data from three hospitals were analyzed.The modeling cohort included data from the First Affiliated Hospital of Zhengzhou University and the People’s Hospital of Xinjiang Uygur Autonomous Region,with Peking University Third Hospital data serving as the external test set.Four machine learning algorithms—logistic regression(LR),multilayer perceptron(MLP),Gaussian naive Bayes(GNB),and random forest(RF)—were used to develop predictive models based on 34 early-accessible clinical variables.A simplifi ed model was then derived based on fi ve key variables(Stanford type,pericardial eff usion,asymmetric peripheral arterial pulsation,decreased bowel sounds,and dyspnea)via Least Absolute Shrinkage and Selection Operator(LASSO)regression to improve ED applicability.RESULTS:A total of 929 patients were included in the modeling cohort,and 210 were included in the external test set.Four machine learning models based on 34 clinical variables were developed,achieving internal and external validation AUCs of 0.85-0.90 and 0.73-0.85,respectively.The simplifi ed model incorporating fi ve key variables demonstrated internal and external validation AUCs of 0.71-0.86 and 0.75-0.78,respectively.Both models showed robust calibration and predictive stability across datasets.CONCLUSION:Both kinds of models were built based on machine learning tools,and proved to have certain prediction performance and extrapolation.展开更多
BACKGROUND Laparoscopic distal pancreatectomy(LDP)has emerged as the preferred approach for both benign and malignant lesions located in the pancreatic body and tail.Nevertheless,a notable deficiency persists in the a...BACKGROUND Laparoscopic distal pancreatectomy(LDP)has emerged as the preferred approach for both benign and malignant lesions located in the pancreatic body and tail.Nevertheless,a notable deficiency persists in the absence of a standardized,procedure-specific metric for evaluating and comparing surgical quality.A composite measure termed“textbook outcome(TO)”,which encompasses key short-term endpoints,has been validated in laparoscopic pancreatoduodenectomy but has not yet been established in dedicated LDP cohorts.The definition and prediction of TO in this context could aid in facilitating cross-institutional benchmarking and fostering advancements in quality improvement.AIM To establish procedure-specific criteria for TO and identify independent predictors of TO failure in patients undergoing LDP.METHODS Consecutive patients who underwent LDP at a single high-volume pancreatic center between January 2015 and August 2022 were retrospectively analyzed.TO was defined as the absence of clinically relevant postoperative pancreatic fistula(grade B/C),post-pancreatectomy hemorrhage(grade B/C),severe complications(Clavien-Dindo≥III),readmission within 30 days,and in-hospital or 30-day mortality.Multivariable logistic regression was employed to identify independent predictors of TO failure,and a nomogram was constructed and internally validated.RESULTS Among 405 eligible patients,286(70.6%)attained TO.Multivariable analysis revealed that female sex[odds ratio(OR)=0.62,95%confidence interval(CI):0.39-0.99]conferred a protective effect,while preoperative endoscopic ultrasound-guided fine-needle aspiration(OR=2.66,95%CI:1.05-6.73),pancreatic portal hypertension(OR=2.81,95%CI:1.06-7.45),and cystic-solid(OR=2.51,95%CI:1.34-4.69)or solid lesions(OR=1.91,95%CI:1.06-3.44)were independently associated with TO failure(all P<0.05).The derived nomogram exhibited modest discrimination and calibration when assessed in both the training and validation datasets.CONCLUSION The proposed LDP-specific definition of TO is feasible and discriminative,and the developed nomogram provides an objective tool for individualized risk assessment.展开更多
Excessive blasting-induced vibration during drilling-and-blasting excavation of deep tunnels can trigger geological hazards and compromise the stability of both the rock mass and support structures.This study focused ...Excessive blasting-induced vibration during drilling-and-blasting excavation of deep tunnels can trigger geological hazards and compromise the stability of both the rock mass and support structures.This study focused on the deep double-line Sejila Mountain tunnel to systematically analyze the spatial response of blasting-induced vibration and to develop a prediction model through field tests and numerical simulations.The results revealed that the presence of a cross passage significantly altered propagation paths and the spatial distribution of blasting-induced vibration velocity.The peak particle velocity(PPV)at the cross-passage corner was amplified by approximately 1.92 times due to wave reflection and geometric focusing.Blasting-induced vibration waves attenuated non-uniformly across the tunnel cross-section,where PPV on the blast-face side was 1.54–6.56 times higher than that on the opposite side.We propose an improved PPV attenuation model that accounts for the propagation path effect.This model significantly improved fitting accuracy and resolved anomalous parameter(k and a)estimates in traditional equations,thereby improving prediction reliability.Furthermore,based on the observed spatial distribution of blasting-induced vibration,optimal monitoring point placement and targeted vibration control measures for tunnel blasting were discussed.These findings provide a scientific basis for designing blasting schemes and vibration mitigation strategies in deep tunnels.展开更多
Current shipping,tourism,and resource development requirements call for more accurate predictions of the Arctic sea-ice concentration(SIC).However,due to the complex physical processes involved,predicting the spatiote...Current shipping,tourism,and resource development requirements call for more accurate predictions of the Arctic sea-ice concentration(SIC).However,due to the complex physical processes involved,predicting the spatiotemporal distribution of Arctic SIC is more challenging than predicting its total extent.In this study,spatiotemporal prediction models for monthly Arctic SIC at 1-to 3-month leads are developed based on U-Net-an effective convolutional deep-learning approach.Based on explicit Arctic sea-ice-atmosphere interactions,11 variables associated with Arctic sea-ice variations are selected as predictors,including observed Arctic SIC,atmospheric,oceanic,and heat flux variables at 1-to 3-month leads.The prediction skills for the monthly Arctic SIC of the test set(from January 2018 to December 2022)are evaluated by examining the mean absolute error(MAE)and binary accuracy(BA).Results showed that the U-Net model had lower MAE and higher BA for Arctic SIC compared to two dynamic climate prediction systems(CFSv2 and NorCPM).By analyzing the relative importance of each predictor,the prediction accuracy relies more on the SIC at the 1-month lead,but on the surface net solar radiation flux at 2-to 3-month leads.However,dynamic models show limited prediction skills for surface net solar radiation flux and other physical processes,especially in autumn.Therefore,the U-Net model can be used to capture the connections among these key physical processes associated with Arctic sea ice and thus offers a significant advantage in predicting Arctic SIC.展开更多
Traditional psychiatric diagnosis relies on subjective symptom assessment,lacking objective biomarkers that hinder early detection and personalized treatment.Plasma proteins and polygenic risk score(PRS),as potential ...Traditional psychiatric diagnosis relies on subjective symptom assessment,lacking objective biomarkers that hinder early detection and personalized treatment.Plasma proteins and polygenic risk score(PRS),as potential predictive tools,hold promise for advancing early diagnosis of mental disorders.This study aims to evaluate the predictive potential of proteomic features and PRS in multiple mental illnesses(depression,schizophrenia,and post-traumatic stress disorder(PTSD)).Using participant data from the UK Biobank-Pharma Proteomics Project,we screen protein associations with mental disorders through least absolute shrinkage and selection operator(LASSO)analysis and construct a Cox regression risk prediction model by integrating the PRS.Additionally,we evaluate predictive performance using 6 machine learning methods and Kaplan-Meier survival curves.Our findings reveal distinct predictive patterns across dis-orders.For depression,integrating plasma proteins with PRS significantly improves prediction beyond the clinical model(C-index=0.6322).For schizophrenia,adding plasma proteins enhances predictive performance,whereas PRS provides no significant improvement.For PTSD,neither plasma proteins nor PRS add substantial predictive value beyond clinical variables.Risk stratification analysis demonstrat that all three mental disorders models can clearly distinguish high-risk from low-risk groups(depression:HR=2.34,P<0.001;schizophrenia:HR=5.47,P<0.001;PTSD:HR=3.02,P<0.001).Al-though it shows good performance in short-term prediction,its long-term prediction ability has decreased,and it needs to be further optimized in the future.This study underscores the differential utility of biomarkers across mental disorders and provides a rationale for disorder-specific predictive modeling in precision psychiatry.展开更多
The micro-riblet structures have been demonstrated effective in controlling the Total Pressure Loss(TPL)of aero-engine blades.However,due to the considerable scale gap between micro-texture and an actual aero-engine b...The micro-riblet structures have been demonstrated effective in controlling the Total Pressure Loss(TPL)of aero-engine blades.However,due to the considerable scale gap between micro-texture and an actual aero-engine blade,wind tunnel tests and numerical simulations with massive grids directly describing the global flow field are costly for aerodynamic evaluation.Furthermore,the fine micro surface structure brings unavoidable manufacturing errors,and the probability prediction contributes to gaining the confidence interval of the results.Therefore,a novel relay-based probabilistic model for multi-fidelity scenarios in the TPL prediction of a compressor cascade with micro-riblet surfaces is proposed to trade off accuracy and efficiency.Combined with the low-fidelity flow data generated by an aerodynamic solution strategy using the boundary surrogate model and the high-fidelity flow data from the experiment,the relay-based modeling has been achieved through knowledge transferring,and the confidence interval can be provided by the Gaussian Process Regression(GPR)model.The TPL of compressor cascades with micro-riblet surfaces under different surface structures at March number Ma=0.64,0.74,0.84 have been evaluated using the Relay-Based Probabilistic(RBP)model.The results illustrate that the RBP model could provide higher accuracy than the Single-Fidelity-Data-Driven(SFDD)prediction model,which show the promising potential of multi-fidelity scenarios data fusion in the aerodynamic evaluation of multi-scale configurations.展开更多
This study traces the development of the Shanghai Typhoon Model(SHTM)from a traditional physics-based regional model toward a data-driven,machine-learning typhoon forecasting system.After upgrading its initial and bou...This study traces the development of the Shanghai Typhoon Model(SHTM)from a traditional physics-based regional model toward a data-driven,machine-learning typhoon forecasting system.After upgrading its initial and boundary conditions,SHTM now leverages large-scale constraints from machine-learning weather prediction(MLWP)models,resulting in an ML–physics hybrid framework.During Typhoon Danas(2025),the hybrid SHTM achieves substantially lower track errors than both the advanced ECMWF Integrated Forecasting System(IFS)and leading MLWP models such as PanGu and FuXi.Furthermore,the hybrid SHTM consistently maintains mean track errors below 200 km up to a forecast lead time of 108 hours,representing a significant advancement in forecast accuracy.In addition,this study highlights the technical roadmap for transitioning from a physics-based typhoon model to a fully data-driven ML typhoon forecast system.It also emphasizes that advances in the physical modeling framework provide a critical foundation for further improving the performance of future data-driven ML typhoon models.展开更多
Background Multibreed genomic prediction(MBGP)is crucial for improving prediction accuracy for breeds with small populations,for which limited data are often available.Recent studies have demonstrated that partitionin...Background Multibreed genomic prediction(MBGP)is crucial for improving prediction accuracy for breeds with small populations,for which limited data are often available.Recent studies have demonstrated that partitioning the genome into nonoverlapping blocks to model heterogeneous genetic(co)variance in multitrait models can achieve higher joint prediction accuracy.However,the block partitioning method,a key factor influencing model performance,has not been extensively explored.Results We introduce mbBayesABLD,a novel Bayesian MBGP model that partitions each chromosome into nonoverlapping blocks on the basis of linkage disequilibrium(LD)patterns.In this model,marker effects within each block are assumed to follow normal distributions with block-specific parameters.We employ simulated data as well as empirical datasets from pigs and beans to assess genomic prediction accuracy across different models using cross-validation.The results demonstrate that mbBayesABLD significantly outperforms conventional MBGP models,such as GBLUP and BayesR.For the meat marbling score trait in pigs,compared with GBLUP,which does not account for heterogeneous genetic(co)variance,mbBayesABLD improves the prediction accuracy for the small-population breed Landrace by 15.6%.Furthermore,our findings indicate that a moderate level of similarity in LD patterns between breeds(with an average correlation of 0.6)is sufficient to improve the prediction accuracy of the target breed.Conclusions This study presents a novel LD block-based approach for multibreed genomic prediction.Our work provides a practical tool for livestock breeding programs and offers new insights into leveraging genetic diversity across breeds for improved genomic prediction.展开更多
Modern industrial environments require uninterrupted machinery operation to maintain productivity standards while ensuring safety and minimizing costs.Conventional maintenance methods,such as reactive maintenance(i.e....Modern industrial environments require uninterrupted machinery operation to maintain productivity standards while ensuring safety and minimizing costs.Conventional maintenance methods,such as reactive maintenance(i.e.,run to failure)or time-based preventive maintenance(i.e.,scheduled servicing),prove ineffective for complex systems with many Internet of Things(IoT)devices and sensors because they fall short in detecting faults at early stages when it is most crucial.This paper presents a predictive maintenance framework based on a hybrid deep learning model that integrates the capabilities of Long Short-Term Memory(LSTM)Networks and Convolutional Neural Networks(CNNs).The framework integrates spatial feature extraction and temporal sequence modeling to accurately classify the health state of industrial equipment into three categories,including Normal,Require Maintenance,and Failed.The framework uses a modular pipeline that includes IoT-enabled data collection along with secure transmission methods to manage cloud storage and provide real-time fault classification.The FD004 subset of the NASA C-MAPSS dataset,containing multivariate sensor readings from aircraft engines,serves as the training and evaluation data for the model.Experimental results show that the LSTM-CNN model outperforms baseline models such as LSTM-SVM and LSTM-RNN,achieving an overall average accuracy of 86.66%,precision of 86.00%,recall of 86.33%,and F1-score of 86.33%.Contrary to the previous LSTM-CNN-based predictive maintenance models that either provide a binary classification or rely on synthetically balanced data,our paper provides a three-class maintenance state(i.e.,Normal,Require Maintenance,and Failed)along with threshold-based labeling that retains the true nature of the degradation.In addition,our work also provides an IoT-to-cloud-based modular architecture for deployment.It offers Computerized Maintenance Management System(CMMS)integration,making our proposed solution not only technically sound but also practical and innovative.The solution achieves real-world industrial deployment readiness through its reliable performance alongside its scalable system design.展开更多
Objective To develop an onset risk prediction nomogram for patients with homocysteine-type(H-type)hypertension(HTH)based on pulse diagram parameters to assist early clinical prediction and diagnosis of HTH.Methods Pat...Objective To develop an onset risk prediction nomogram for patients with homocysteine-type(H-type)hypertension(HTH)based on pulse diagram parameters to assist early clinical prediction and diagnosis of HTH.Methods Patients diagnosed with essential hypertension and admitted to Shanghai Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine,Shang-hai Hospital of Traditional Chinese Medicine,and Shanghai Hospital of Integrated Tradition-al Chinese and Western Medicine from July 6th 2020 to June 16th 2021,and from August 11th 2023 to January 22nd 2024,were enrolled in this retrospective research.The baselines and clinical biochemical indicators of patients were collected.The SMART-I TCM pulse instru-ment was applied to gather pulse diagram parameters.Multivariate logistic regression was adopted to analyze the risk factors for HTH.RStudio was employed to construct the nomo-gram model,receiver operating characteristic(ROC)curve,and calibration curve(bootstrap self-sampling 200 times),and clinical decision curve were drawn to evaluate the model’s dis-crimination and clinical effectiveness.Results A total of 168 hospitalized patients with essential hypertension were selected and di-vided into non-HTH group(n=29)and HTH group(n=139).Compared with non-HTH group,HTH group had a lower body mass index(BMI),and higher proportions of male pa-tients and drinkers(P<0.05).The ventricular wall thickening(VWT)could not be deter-mined.The proportions of left common carotid intima-media wall thickness(LCCIMWT)and serum creatinine(SCR)were higher in HTH group(P<0.05).The pulse diagram parameter As was significantly higher,and H4/H1 and T1/T were lower in HTH group(P<0.05).Gender,al-cohol consumption,serum creatinine,and the pulse diagram parameter H4/H1 were identi-fied as independent risk factors for HTH(P<0.05).The nomogram’s area under the ROC curve(AUC)was 0.795[95%confidence interval(CI):(0.7066,0.8828)],with a specificity of 0.724 and sensitivity of 0.799.After 200 times repeated bootstrap self-samplings,the calibra-tion curve showed that the simulated curve fits well with the actual curve(x^(2)=9.5002,P=0.3019).The clinical decision curve indicated that the nomogram’s applicability was optimal when the threshold for predicting HTH was between 0.38 and 1.00.Conclusion The nomogram model could be valuable for predicting the onset risk of HTH and pulse diagram parameters can facilitate early screening and prevention of HTH.展开更多
Predicting cross-immunity between viral strains is vital for public health surveillance and vaccine development.Traditional neural network methods,such as BiLSTM,could be ineffective due to the lack of lab data for mo...Predicting cross-immunity between viral strains is vital for public health surveillance and vaccine development.Traditional neural network methods,such as BiLSTM,could be ineffective due to the lack of lab data for model training and the overshadowing of crucial features within sequence concatenation.The current work proposes a less data-consuming model incorporating a pre-trained gene sequence model and a mutual information inference operator.Our methodology utilizes gene alignment and deduplication algorithms to preprocess gene sequences,enhancing the model’s capacity to discern and focus on distinctions among input gene pairs.The model,i.e.,DNA Pretrained Cross-Immunity Protection Inference model(DPCIPI),outperforms state-of-theart(SOTA)models in predicting hemagglutination inhibition titer from influenza viral gene sequences only.Improvement in binary cross-immunity prediction is 1.58%in F1,2.34%in precision,1.57%in recall,and 1.57%in Accuracy.For multilevel cross-immunity improvements,the improvement is 2.12%in F1,3.50%in precision,2.19%in recall,and 2.19%in Accuracy.Our study showcases the potential of pre-trained gene models to improve predictions of antigenic variation and cross-immunity.With expanding gene data and advancements in pre-trained models,this approach promises significant impacts on vaccine development and public health.展开更多
Improving the efficiency of athletic performance and reducing the likelihood of overtraining are primarily determined goals that can be achieved by the correct organization of the training process.The nature of adapta...Improving the efficiency of athletic performance and reducing the likelihood of overtraining are primarily determined goals that can be achieved by the correct organization of the training process.The nature of adaptation to physical stress is associated with the specificity,focus,and degree of biochemical and functional changes that occur during muscular work.In this study,we aimed to develop a diagnostic model for predicting metabolic processes in athletes based on standard biochemical blood analysis indicators.The study involved athletes from the track and field athletics team(men,n=42,average age was[22.55±3.68]years).Blood samples were collected in the morning at the beginning and end of the training week during the annual cycle.During the entire period,3625 laboratory parameter tests were conducted.Capillary blood sampling in athletes was conducted from the distal phalanx of the finger after overnight fasting,according to standard diagnostic procedures.To determine the predominance of anabolic or catabolic processes,equations were derived from a linear discriminant function.The discriminant function of predicting metabolic processes in athletes has a high information capacity(92.1%),as confirmed by the biochemical results of neuroendocrine system activity,which characterized the body's stage of adaptive regulatory mechanisms in response to stress factors.The classification matrix used to predict the metabolic processes based on the results of the discriminant function calculation demonstrates the statistical significance of the model(p<0.01).Consequently,an informative mathematical model was developed,which enabled the reliable and timely prediction of the prevalence of one of the metabolic activity phases in the athlete's body.The use of the developed model will also allow us to assess the nature of adaptation to specific muscular work,identify an athlete's weaknesses,forecast the success of their performance,and timely adjust both the training process and the recovery program.展开更多
In this study,a machine learning-based predictive model was developed for the Musa petti Wind Farm in Sri Lanka to address the need for localized forecasting solutions.Using data on wind speed,air temperature,nacelle ...In this study,a machine learning-based predictive model was developed for the Musa petti Wind Farm in Sri Lanka to address the need for localized forecasting solutions.Using data on wind speed,air temperature,nacelle position,and actual power,lagged features were generated to capture temporal dependencies.Among 24 evaluated models,the ensemble bagging approach achieved the best performance,with R^(2) values of 0.89 at 0 min and 0.75 at 60 min.Shapley Additive exPlanations(SHAP)analysis revealed that while wind speed is the primary driver for short-term predictions,air temperature and nacelle position become more influential at longer forecasting horizons.These findings underscore the reliability of short-term predictions and the potential benefits of integrating hybrid AI and probabilistic models for extended forecasts.Our work contributes a robust and explainable framework to support Sri Lanka’s renewable energy transition,and future research will focus on real-time deployment and uncertainty quantification.展开更多
基金supported by the National Natural Science Foundation of China(Nos.52278461,52308467,and 52425213).
文摘The uneven distribution of the temperature field in the track structure,caused by various meteorological factors such as extremely low temperatures and snowfall,leads to significant temperature loads and is the primary cause of damage to China Railway Track System(CRTS)III ballastless tracks in cold regions during service.In this study,to predict the temperature of the track structure accurately,we analyzed meteorological data collected from Shenyang,China,and identified the factors that had the most effect on the track temperature field.We propose a temporal convolutional network(TCN)-based temperature field prediction model for ballastless tracks(TCN-Track model),which enhances the ability to extract and fuse local and global features from complex long-term meteorological data.The results indicate that the proposed TCN-Track model performs well in predicting track temperature fields from meteorological data,with a mean absolute error(MAE)ranging from 0.26 to 0.39,a root mean square error(RMSE)ranging from 0.32 to 0.50,and correlation coefficient(R)values ranging from 0.888 to 0.985.Compared with a long short-term memory(LSTM)model,the MAE of the TCN-Track model is reduced by 89.17%and the RMSE by 88.51%.This method offers a new solution for accurately predicting the temperature field of ballastless tracks in cold regions,aiding in predicting and preventing track damage caused by low temperatures.
基金supported by the National Key Research and Development Program of China(Grant No.2022YFC3002803)the National Key Research and Development Program of China(Grant No.2024YFF0808402)the National Natural Science Foundation of China(Grant No.42375169)。
文摘Northeast China serves as an important crop production region.Accurately forecasting summer precipitation in Northeast China(NEC-PR)has been a challenge due to its wide range of time scales influenced by varying climatic conditions.This study presents a scale separation hybrid statistical model with recurrent neural network(SS-RNN)to predict the summer monthly NEC-PR.The SS-RNN model decomposes the multiple scales of the NEC-PR into several spatiotemporal intrinsic mode functions covering annual to decadal time scales.This strategy provides a way to derive appropriate predictors and establish predictive models for the primary spatial modes of the NEC-PR at various time scales.Our results demonstrate substantial improvements by the SS-RNN model in predicting the summer monthly NEC-PR as compared with dynamic models,particularly in predicting the spatial pattern of the NEC-PR.In this paper we take August,the month of the highest NEC-PR,to assess our model skill.Independent forecasts of the August NEC-PR over the period 2021–24 achieve significant spatial anomaly correlation coefficients,reaching a maximum value of 0.83.Additional verifications by station observations show that the model hits most station anomalies,achieving a mean predictive skill score of 90.
基金Funded by State Railway Administration Research Project(No.2023JS007)National Natural Science Foundation of China(No.52438002)+1 种基金Research and Development Programs for Science and Technology of China Railways Corporation(No.J2023G003)New Cornerstone Science Foundation through the XPLORER PRIZE。
文摘To investigate the influence of coarse aggregate parent rock properties on the elastic modulus of concrete,the mineralogical properties and stress-strain curves of granite and dolomite parent rocks,as well as the strength and elastic modulus of mortar and concrete prepared with mechanism aggregates of the corresponding lithology,and the stress-strain curves of concrete were investigated.In this paper,a coarse aggregate and mortar matrix bonding assumption is proposed,and a prediction model for the elastic modulus of mortar is established by considering the lithology of the mechanism sand and the slurry components.An equivalent coarse aggregate elastic modulus model was established by considering factors such as coarse aggregate particle size,volume fraction,and mortar thickness between coarse aggregates.Based on the elastic modulus of the equivalent coarse aggregate and the remaining mortar,a prediction model for the elastic modulus of the two and three components of concrete in series and then in parallel was established,and the predicted values differed from the measured values within 10%.It is proposed that the coarse aggregate elastic modulus in highstrength concrete is the most critical factor affecting the elastic modulus of concrete,and as the coarse aggregate elastic modulus increases by 27.7%,the concrete elastic modulus increases by 19.5%.
文摘Customer churn is the rate at which customers discontinue doing business with a company over a given time period.It is an essential measure for businesses to monitor high churn rates,as they often indicate underlying issues with services,products,or customer experience,resulting in considerable income loss.Prediction of customer churn is a crucial task aimed at retaining customers and maintaining revenue growth.Traditional machine learning(ML)models often struggle to capture complex temporal dependencies in client behavior data.To address this,an optimized deep learning(DL)approach using a Regularized Bidirectional Long Short-Term Memory(RBiLSTM)model is proposed to mitigate overfitting and improve generalization error.The model integrates dropout,L2-regularization,and early stopping to enhance predictive accuracy while preventing over-reliance on specific patterns.Moreover,this study investigates the effect of optimization techniques on boosting the training efficiency of the developed model.Experimental results on a recent public customer churn dataset demonstrate that the trained model outperforms the traditional ML models and some other DL models,such as Long Short-Term Memory(LSTM)and Deep Neural Network(DNN),in churn prediction performance and stability.The proposed approach achieves 96.1%accuracy,compared with LSTM and DNN,which attain 94.5%and 94.1%accuracy,respectively.These results confirm that the proposed approach can be used as a valuable tool for businesses to identify at-risk consumers proactively and implement targeted retention strategies.
基金funded by Scion's Strategic Science Investment Fund(SSIF)the Forest Growers Levy Trust(FGLT)through the Resilient Forests Programme(Task No.A89220)。
文摘Pinus radiata(D.Don)dominates New Zealand's forestry industry,constituting 91%of plantations,and is among the world's most important plantation species.Given the socio-economic and environmental importance of this species,it is important to have accurate and precise projections over time to make efficient decisions for forest management and greenfield investments in afforestation projects,especially for permanent carbon forests.Future projections of any natural resource systems rely on modeling;however,the acceleration of climate change makes future projections of yield less certain.These challenges also impact national expectations of the contribution planted forests will provide to address climate change and meet international commitments under the Paris Agreement.Using a large national-scale set of contemporary ground-measured data(2013–2023),this study investigates the performance of two growth models developed over 30 years ago that are widely used by NZ plantation growers:1)the Pumice Plateau Model 1988(PPM88)and 2)the 300-index(including a model variant of regional drift).Model simulations were made using the FORECASTER modeling suite with geographic boundaries to adjust for drift in space and time.Basal area(BA,m^(2)⋅ha^(-1))and volume(m^(3)⋅ha^(-1))were simulated,and standard errors and goodness-of-fit metrics calculated up to a typical rotation age of 30 years.Model residuals were then separated and analysed for the main plantation growing regions.The models overpredicted observed growth by between 6.8%and 16.2%,but model predictions and errors varied significantly between regions.The results of this study provided clear evidence of divergence between the outputs of both models and the measured data.Finally,this study suggests future measures to address challenges posed by these discrepancies that will provide better information for forest management and investment decisions in a changing climate.
基金Construction and Application of Frailty Trajectory Prediction Model for Middle-aged and Elderly Patients with Upper Gastrointestinal Bleeding,Project Source:Sichuan Vocational College of Nursing(Project No.:2024ZRY25)。
文摘Objective:To investigate the impact of targeted nursing interventions based on frailty prediction models on peri-hospitalization clinical outcomes in middle-aged and elderly patients with upper gastrointestinal bleeding(UGIB).Methods:A prospective cohort study was conducted,and 126 middle-aged and elderly patients with UGIB admitted from August 2024 to August 2025 were selected as the study subjects.The patients were divided into the intervention group(63 cases)and the control group(63 cases)based on whether they received nursing intervention based on frailty prediction models.The control group received routine care,while the intervention group,on the basis of routine care,used the FRAIL scale combined with laboratory indicators(albumin,hemoglobin,etc.)to establish a predictive model to evaluate patients within 24 hours of admission,and implemented multi-dimensional targeted nursing intervention for pre-frailty or frailty patients screened out.The incidence of frailty,rebleeding rate,average length of stay,hospitalization cost,and nursing satisfaction during hospitalization were compared between the two groups.Results:The incidence of frailty during hospitalization in the intervention group was 11.1%(7 cases/63 cases),significantly lower than 31.7%(20 cases/63 cases)in the control group,and the difference was statistically significant(p<0.05).The rebleeding rate of 4.8%vs 12.7%,the average length of stay of(7.2±1.5)days vs(9.1±2.2)days,and the average hospitalization cost of(23,000±6,000)yuan vs(28,000±7,000)yuan in the intervention group were all lower than those in the control group(all p<0.05).The nursing satisfaction score of the intervention group(93.5±4.2)points was higher than that of the control group(86.3±5.8)points(p<0.05).Conclusion:The frailty prediction model applied to the peri-hospitalization care of middle-aged and elderly patients with UGIB can effectively identify frailty risk.Through early targeted intervention,the incidence of frailty and rebleeding rate can be reduced,the length of hospital stay can be shortened,medical expenses can be reduced,and nursing satisfaction can be improved,which has clinical promotion value.
基金financially supported by the National Key Research and Development Program of China (No. 2023YFB3812601)the National Natural Science Foundation of China (No. 51925401)the Young Elite Scientists Sponsorship Program by CAST, China (No. 2022QNRC001)。
文摘Machine learning-assisted methods for rapid and accurate prediction of temperature field,mushy zone,and grain size were proposed for the heating−cooling combined mold(HCCM)horizontal continuous casting of C70250 alloy plates.First,finite element simulations of casting processes were carried out with various parameters to build a dataset.Subsequently,different machine learning algorithms were employed to achieve high precision in predicting temperature fields,mushy zone locations,mushy zone inclination angle,and billet grain size.Finally,the process parameters were quickly optimized using a strategy consisting of random generation,prediction,and screening,allowing the mushy zone to be controlled to the desired target.The optimized parameters are 1234℃for heating mold temperature,47 mm/min for casting speed,and 10 L/min for cooling water flow rate.The optimized mushy zone is located in the middle of the second heat insulation section and has an inclination angle of roughly 7°.
基金supported by the special fund of the National Clinical Key Specialty Construction Program[(2022)301-2305].
文摘BACKGROUND:This study aims to develop and validate a machine learning-based in-hospital mortality predictive model for acute aortic syndrome(AAS)in the emergency department(ED)and to derive a simplifi ed version suitable for rapid clinical application.METHODS:In this multi-center retrospective cohort study,AAS patient data from three hospitals were analyzed.The modeling cohort included data from the First Affiliated Hospital of Zhengzhou University and the People’s Hospital of Xinjiang Uygur Autonomous Region,with Peking University Third Hospital data serving as the external test set.Four machine learning algorithms—logistic regression(LR),multilayer perceptron(MLP),Gaussian naive Bayes(GNB),and random forest(RF)—were used to develop predictive models based on 34 early-accessible clinical variables.A simplifi ed model was then derived based on fi ve key variables(Stanford type,pericardial eff usion,asymmetric peripheral arterial pulsation,decreased bowel sounds,and dyspnea)via Least Absolute Shrinkage and Selection Operator(LASSO)regression to improve ED applicability.RESULTS:A total of 929 patients were included in the modeling cohort,and 210 were included in the external test set.Four machine learning models based on 34 clinical variables were developed,achieving internal and external validation AUCs of 0.85-0.90 and 0.73-0.85,respectively.The simplifi ed model incorporating fi ve key variables demonstrated internal and external validation AUCs of 0.71-0.86 and 0.75-0.78,respectively.Both models showed robust calibration and predictive stability across datasets.CONCLUSION:Both kinds of models were built based on machine learning tools,and proved to have certain prediction performance and extrapolation.
文摘BACKGROUND Laparoscopic distal pancreatectomy(LDP)has emerged as the preferred approach for both benign and malignant lesions located in the pancreatic body and tail.Nevertheless,a notable deficiency persists in the absence of a standardized,procedure-specific metric for evaluating and comparing surgical quality.A composite measure termed“textbook outcome(TO)”,which encompasses key short-term endpoints,has been validated in laparoscopic pancreatoduodenectomy but has not yet been established in dedicated LDP cohorts.The definition and prediction of TO in this context could aid in facilitating cross-institutional benchmarking and fostering advancements in quality improvement.AIM To establish procedure-specific criteria for TO and identify independent predictors of TO failure in patients undergoing LDP.METHODS Consecutive patients who underwent LDP at a single high-volume pancreatic center between January 2015 and August 2022 were retrospectively analyzed.TO was defined as the absence of clinically relevant postoperative pancreatic fistula(grade B/C),post-pancreatectomy hemorrhage(grade B/C),severe complications(Clavien-Dindo≥III),readmission within 30 days,and in-hospital or 30-day mortality.Multivariable logistic regression was employed to identify independent predictors of TO failure,and a nomogram was constructed and internally validated.RESULTS Among 405 eligible patients,286(70.6%)attained TO.Multivariable analysis revealed that female sex[odds ratio(OR)=0.62,95%confidence interval(CI):0.39-0.99]conferred a protective effect,while preoperative endoscopic ultrasound-guided fine-needle aspiration(OR=2.66,95%CI:1.05-6.73),pancreatic portal hypertension(OR=2.81,95%CI:1.06-7.45),and cystic-solid(OR=2.51,95%CI:1.34-4.69)or solid lesions(OR=1.91,95%CI:1.06-3.44)were independently associated with TO failure(all P<0.05).The derived nomogram exhibited modest discrimination and calibration when assessed in both the training and validation datasets.CONCLUSION The proposed LDP-specific definition of TO is feasible and discriminative,and the developed nomogram provides an objective tool for individualized risk assessment.
基金financially supported by the National Natural Science Foundation of China(Nos.42577209 and U22A20239)the Key R&D Program of Hunan Province(No.2024WK2004)the Key Technologies for Accurate Diagnosis and Intelligent Prevention and Control of Slope Hazards in Open pit Mines,181 Major R&D projects of Metallurgical Corporation of China Ltd。
文摘Excessive blasting-induced vibration during drilling-and-blasting excavation of deep tunnels can trigger geological hazards and compromise the stability of both the rock mass and support structures.This study focused on the deep double-line Sejila Mountain tunnel to systematically analyze the spatial response of blasting-induced vibration and to develop a prediction model through field tests and numerical simulations.The results revealed that the presence of a cross passage significantly altered propagation paths and the spatial distribution of blasting-induced vibration velocity.The peak particle velocity(PPV)at the cross-passage corner was amplified by approximately 1.92 times due to wave reflection and geometric focusing.Blasting-induced vibration waves attenuated non-uniformly across the tunnel cross-section,where PPV on the blast-face side was 1.54–6.56 times higher than that on the opposite side.We propose an improved PPV attenuation model that accounts for the propagation path effect.This model significantly improved fitting accuracy and resolved anomalous parameter(k and a)estimates in traditional equations,thereby improving prediction reliability.Furthermore,based on the observed spatial distribution of blasting-induced vibration,optimal monitoring point placement and targeted vibration control measures for tunnel blasting were discussed.These findings provide a scientific basis for designing blasting schemes and vibration mitigation strategies in deep tunnels.
基金supported by the National Key Research and Development Program of China[grant number 2022YFE0106800]an Innovation Group Project of the Southern Marine Science and Engineering Guangdong Laboratory(Zhuhai)[grant number 311024001]+3 种基金a project supported by the Southern Marine Science and Engineering Guangdong Laboratory(Zhuhai)[grant number SML2023SP209]a Research Council of Norway funded project(MAPARC)[grant number 328943]a Nansen Center´s basic institutional funding[grant number 342624]the high-performance computing support from the School of Atmospheric Science at Sun Yat-sen University。
文摘Current shipping,tourism,and resource development requirements call for more accurate predictions of the Arctic sea-ice concentration(SIC).However,due to the complex physical processes involved,predicting the spatiotemporal distribution of Arctic SIC is more challenging than predicting its total extent.In this study,spatiotemporal prediction models for monthly Arctic SIC at 1-to 3-month leads are developed based on U-Net-an effective convolutional deep-learning approach.Based on explicit Arctic sea-ice-atmosphere interactions,11 variables associated with Arctic sea-ice variations are selected as predictors,including observed Arctic SIC,atmospheric,oceanic,and heat flux variables at 1-to 3-month leads.The prediction skills for the monthly Arctic SIC of the test set(from January 2018 to December 2022)are evaluated by examining the mean absolute error(MAE)and binary accuracy(BA).Results showed that the U-Net model had lower MAE and higher BA for Arctic SIC compared to two dynamic climate prediction systems(CFSv2 and NorCPM).By analyzing the relative importance of each predictor,the prediction accuracy relies more on the SIC at the 1-month lead,but on the surface net solar radiation flux at 2-to 3-month leads.However,dynamic models show limited prediction skills for surface net solar radiation flux and other physical processes,especially in autumn.Therefore,the U-Net model can be used to capture the connections among these key physical processes associated with Arctic sea ice and thus offers a significant advantage in predicting Arctic SIC.
基金The National Natural Science Foundation of China-Regional Science“Identification of novel drug targets for lung cancer via Mendelian randomization analysis based on blood proteomics”(62362062)The 2025 Xinjiang University Excellent Graduate Innovation Project“Research on identification of therapeutic targets and predictive factors for mental disorders based on proteomics”(XJDX2025YJS151)。
文摘Traditional psychiatric diagnosis relies on subjective symptom assessment,lacking objective biomarkers that hinder early detection and personalized treatment.Plasma proteins and polygenic risk score(PRS),as potential predictive tools,hold promise for advancing early diagnosis of mental disorders.This study aims to evaluate the predictive potential of proteomic features and PRS in multiple mental illnesses(depression,schizophrenia,and post-traumatic stress disorder(PTSD)).Using participant data from the UK Biobank-Pharma Proteomics Project,we screen protein associations with mental disorders through least absolute shrinkage and selection operator(LASSO)analysis and construct a Cox regression risk prediction model by integrating the PRS.Additionally,we evaluate predictive performance using 6 machine learning methods and Kaplan-Meier survival curves.Our findings reveal distinct predictive patterns across dis-orders.For depression,integrating plasma proteins with PRS significantly improves prediction beyond the clinical model(C-index=0.6322).For schizophrenia,adding plasma proteins enhances predictive performance,whereas PRS provides no significant improvement.For PTSD,neither plasma proteins nor PRS add substantial predictive value beyond clinical variables.Risk stratification analysis demonstrat that all three mental disorders models can clearly distinguish high-risk from low-risk groups(depression:HR=2.34,P<0.001;schizophrenia:HR=5.47,P<0.001;PTSD:HR=3.02,P<0.001).Al-though it shows good performance in short-term prediction,its long-term prediction ability has decreased,and it needs to be further optimized in the future.This study underscores the differential utility of biomarkers across mental disorders and provides a rationale for disorder-specific predictive modeling in precision psychiatry.
基金supported by the National Natural Science Foundation of China(No.12301672)the Shanghai Science and Technology Innovation Action Plan(Yangfan Special Project),China(No.23YF1401300)。
文摘The micro-riblet structures have been demonstrated effective in controlling the Total Pressure Loss(TPL)of aero-engine blades.However,due to the considerable scale gap between micro-texture and an actual aero-engine blade,wind tunnel tests and numerical simulations with massive grids directly describing the global flow field are costly for aerodynamic evaluation.Furthermore,the fine micro surface structure brings unavoidable manufacturing errors,and the probability prediction contributes to gaining the confidence interval of the results.Therefore,a novel relay-based probabilistic model for multi-fidelity scenarios in the TPL prediction of a compressor cascade with micro-riblet surfaces is proposed to trade off accuracy and efficiency.Combined with the low-fidelity flow data generated by an aerodynamic solution strategy using the boundary surrogate model and the high-fidelity flow data from the experiment,the relay-based modeling has been achieved through knowledge transferring,and the confidence interval can be provided by the Gaussian Process Regression(GPR)model.The TPL of compressor cascades with micro-riblet surfaces under different surface structures at March number Ma=0.64,0.74,0.84 have been evaluated using the Relay-Based Probabilistic(RBP)model.The results illustrate that the RBP model could provide higher accuracy than the Single-Fidelity-Data-Driven(SFDD)prediction model,which show the promising potential of multi-fidelity scenarios data fusion in the aerodynamic evaluation of multi-scale configurations.
基金supported by the Special Project-Original Exploration(Grant No.42450163)the National Youth Science Foundation of China Project(Grant No.4240050560)the Research and Development of Key Technologies for Artificial Intelligence Regional Typhoon Forecasting Model project.
文摘This study traces the development of the Shanghai Typhoon Model(SHTM)from a traditional physics-based regional model toward a data-driven,machine-learning typhoon forecasting system.After upgrading its initial and boundary conditions,SHTM now leverages large-scale constraints from machine-learning weather prediction(MLWP)models,resulting in an ML–physics hybrid framework.During Typhoon Danas(2025),the hybrid SHTM achieves substantially lower track errors than both the advanced ECMWF Integrated Forecasting System(IFS)and leading MLWP models such as PanGu and FuXi.Furthermore,the hybrid SHTM consistently maintains mean track errors below 200 km up to a forecast lead time of 108 hours,representing a significant advancement in forecast accuracy.In addition,this study highlights the technical roadmap for transitioning from a physics-based typhoon model to a fully data-driven ML typhoon forecast system.It also emphasizes that advances in the physical modeling framework provide a critical foundation for further improving the performance of future data-driven ML typhoon models.
基金supported by the Biological Breeding-Major Projects in National Science and Technology(No.2023ZD0404405)the Earmarked Fund for China Agriculture Research System(No.CARS-pig-35)+2 种基金the National Natural Science Foundation of China(No.3227284,32302708)the 2115 Talent Development Program of China Agricultural University,the Chinese Universities Scientific Fund(No.2023TC196)the Seed Industry Revitalization Action Project of Guangdong Province(No.2024-XPY-06-001)。
文摘Background Multibreed genomic prediction(MBGP)is crucial for improving prediction accuracy for breeds with small populations,for which limited data are often available.Recent studies have demonstrated that partitioning the genome into nonoverlapping blocks to model heterogeneous genetic(co)variance in multitrait models can achieve higher joint prediction accuracy.However,the block partitioning method,a key factor influencing model performance,has not been extensively explored.Results We introduce mbBayesABLD,a novel Bayesian MBGP model that partitions each chromosome into nonoverlapping blocks on the basis of linkage disequilibrium(LD)patterns.In this model,marker effects within each block are assumed to follow normal distributions with block-specific parameters.We employ simulated data as well as empirical datasets from pigs and beans to assess genomic prediction accuracy across different models using cross-validation.The results demonstrate that mbBayesABLD significantly outperforms conventional MBGP models,such as GBLUP and BayesR.For the meat marbling score trait in pigs,compared with GBLUP,which does not account for heterogeneous genetic(co)variance,mbBayesABLD improves the prediction accuracy for the small-population breed Landrace by 15.6%.Furthermore,our findings indicate that a moderate level of similarity in LD patterns between breeds(with an average correlation of 0.6)is sufficient to improve the prediction accuracy of the target breed.Conclusions This study presents a novel LD block-based approach for multibreed genomic prediction.Our work provides a practical tool for livestock breeding programs and offers new insights into leveraging genetic diversity across breeds for improved genomic prediction.
文摘Modern industrial environments require uninterrupted machinery operation to maintain productivity standards while ensuring safety and minimizing costs.Conventional maintenance methods,such as reactive maintenance(i.e.,run to failure)or time-based preventive maintenance(i.e.,scheduled servicing),prove ineffective for complex systems with many Internet of Things(IoT)devices and sensors because they fall short in detecting faults at early stages when it is most crucial.This paper presents a predictive maintenance framework based on a hybrid deep learning model that integrates the capabilities of Long Short-Term Memory(LSTM)Networks and Convolutional Neural Networks(CNNs).The framework integrates spatial feature extraction and temporal sequence modeling to accurately classify the health state of industrial equipment into three categories,including Normal,Require Maintenance,and Failed.The framework uses a modular pipeline that includes IoT-enabled data collection along with secure transmission methods to manage cloud storage and provide real-time fault classification.The FD004 subset of the NASA C-MAPSS dataset,containing multivariate sensor readings from aircraft engines,serves as the training and evaluation data for the model.Experimental results show that the LSTM-CNN model outperforms baseline models such as LSTM-SVM and LSTM-RNN,achieving an overall average accuracy of 86.66%,precision of 86.00%,recall of 86.33%,and F1-score of 86.33%.Contrary to the previous LSTM-CNN-based predictive maintenance models that either provide a binary classification or rely on synthetically balanced data,our paper provides a three-class maintenance state(i.e.,Normal,Require Maintenance,and Failed)along with threshold-based labeling that retains the true nature of the degradation.In addition,our work also provides an IoT-to-cloud-based modular architecture for deployment.It offers Computerized Maintenance Management System(CMMS)integration,making our proposed solution not only technically sound but also practical and innovative.The solution achieves real-world industrial deployment readiness through its reliable performance alongside its scalable system design.
基金National Natural Science Foundation of China (81973749 and 8143594)State Administration of Traditional Chinese Medicine High-level Chinese Medicine Key Discipline Construction Project (zyyzdxk-2023069)。
文摘Objective To develop an onset risk prediction nomogram for patients with homocysteine-type(H-type)hypertension(HTH)based on pulse diagram parameters to assist early clinical prediction and diagnosis of HTH.Methods Patients diagnosed with essential hypertension and admitted to Shanghai Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine,Shang-hai Hospital of Traditional Chinese Medicine,and Shanghai Hospital of Integrated Tradition-al Chinese and Western Medicine from July 6th 2020 to June 16th 2021,and from August 11th 2023 to January 22nd 2024,were enrolled in this retrospective research.The baselines and clinical biochemical indicators of patients were collected.The SMART-I TCM pulse instru-ment was applied to gather pulse diagram parameters.Multivariate logistic regression was adopted to analyze the risk factors for HTH.RStudio was employed to construct the nomo-gram model,receiver operating characteristic(ROC)curve,and calibration curve(bootstrap self-sampling 200 times),and clinical decision curve were drawn to evaluate the model’s dis-crimination and clinical effectiveness.Results A total of 168 hospitalized patients with essential hypertension were selected and di-vided into non-HTH group(n=29)and HTH group(n=139).Compared with non-HTH group,HTH group had a lower body mass index(BMI),and higher proportions of male pa-tients and drinkers(P<0.05).The ventricular wall thickening(VWT)could not be deter-mined.The proportions of left common carotid intima-media wall thickness(LCCIMWT)and serum creatinine(SCR)were higher in HTH group(P<0.05).The pulse diagram parameter As was significantly higher,and H4/H1 and T1/T were lower in HTH group(P<0.05).Gender,al-cohol consumption,serum creatinine,and the pulse diagram parameter H4/H1 were identi-fied as independent risk factors for HTH(P<0.05).The nomogram’s area under the ROC curve(AUC)was 0.795[95%confidence interval(CI):(0.7066,0.8828)],with a specificity of 0.724 and sensitivity of 0.799.After 200 times repeated bootstrap self-samplings,the calibra-tion curve showed that the simulated curve fits well with the actual curve(x^(2)=9.5002,P=0.3019).The clinical decision curve indicated that the nomogram’s applicability was optimal when the threshold for predicting HTH was between 0.38 and 1.00.Conclusion The nomogram model could be valuable for predicting the onset risk of HTH and pulse diagram parameters can facilitate early screening and prevention of HTH.
基金supported by the Bill & Melinda Gates Foundation and the Minderoo Foundation
文摘Predicting cross-immunity between viral strains is vital for public health surveillance and vaccine development.Traditional neural network methods,such as BiLSTM,could be ineffective due to the lack of lab data for model training and the overshadowing of crucial features within sequence concatenation.The current work proposes a less data-consuming model incorporating a pre-trained gene sequence model and a mutual information inference operator.Our methodology utilizes gene alignment and deduplication algorithms to preprocess gene sequences,enhancing the model’s capacity to discern and focus on distinctions among input gene pairs.The model,i.e.,DNA Pretrained Cross-Immunity Protection Inference model(DPCIPI),outperforms state-of-theart(SOTA)models in predicting hemagglutination inhibition titer from influenza viral gene sequences only.Improvement in binary cross-immunity prediction is 1.58%in F1,2.34%in precision,1.57%in recall,and 1.57%in Accuracy.For multilevel cross-immunity improvements,the improvement is 2.12%in F1,3.50%in precision,2.19%in recall,and 2.19%in Accuracy.Our study showcases the potential of pre-trained gene models to improve predictions of antigenic variation and cross-immunity.With expanding gene data and advancements in pre-trained models,this approach promises significant impacts on vaccine development and public health.
基金financed by the Ministry of Science and Higher Education of the Russian Federation within the framework of state support for the creation and development of World-Class Research Centers‘Digital Biodesign and Personalized Healthcare’No 75-15-2022-305.
文摘Improving the efficiency of athletic performance and reducing the likelihood of overtraining are primarily determined goals that can be achieved by the correct organization of the training process.The nature of adaptation to physical stress is associated with the specificity,focus,and degree of biochemical and functional changes that occur during muscular work.In this study,we aimed to develop a diagnostic model for predicting metabolic processes in athletes based on standard biochemical blood analysis indicators.The study involved athletes from the track and field athletics team(men,n=42,average age was[22.55±3.68]years).Blood samples were collected in the morning at the beginning and end of the training week during the annual cycle.During the entire period,3625 laboratory parameter tests were conducted.Capillary blood sampling in athletes was conducted from the distal phalanx of the finger after overnight fasting,according to standard diagnostic procedures.To determine the predominance of anabolic or catabolic processes,equations were derived from a linear discriminant function.The discriminant function of predicting metabolic processes in athletes has a high information capacity(92.1%),as confirmed by the biochemical results of neuroendocrine system activity,which characterized the body's stage of adaptive regulatory mechanisms in response to stress factors.The classification matrix used to predict the metabolic processes based on the results of the discriminant function calculation demonstrates the statistical significance of the model(p<0.01).Consequently,an informative mathematical model was developed,which enabled the reliable and timely prediction of the prevalence of one of the metabolic activity phases in the athlete's body.The use of the developed model will also allow us to assess the nature of adaptation to specific muscular work,identify an athlete's weaknesses,forecast the success of their performance,and timely adjust both the training process and the recovery program.
文摘In this study,a machine learning-based predictive model was developed for the Musa petti Wind Farm in Sri Lanka to address the need for localized forecasting solutions.Using data on wind speed,air temperature,nacelle position,and actual power,lagged features were generated to capture temporal dependencies.Among 24 evaluated models,the ensemble bagging approach achieved the best performance,with R^(2) values of 0.89 at 0 min and 0.75 at 60 min.Shapley Additive exPlanations(SHAP)analysis revealed that while wind speed is the primary driver for short-term predictions,air temperature and nacelle position become more influential at longer forecasting horizons.These findings underscore the reliability of short-term predictions and the potential benefits of integrating hybrid AI and probabilistic models for extended forecasts.Our work contributes a robust and explainable framework to support Sri Lanka’s renewable energy transition,and future research will focus on real-time deployment and uncertainty quantification.
