Objective The Asia-Pacific region has a high chronic obstructive pulmonary disease(COPD)burden,but studies on its trends are limited.Using the Global Burden of Disease(GBD)2019 data,we analyzed COPD trends in 36 count...Objective The Asia-Pacific region has a high chronic obstructive pulmonary disease(COPD)burden,but studies on its trends are limited.Using the Global Burden of Disease(GBD)2019 data,we analyzed COPD trends in 36 countries and territories from 1990 to 2019 and predicted future incidence trends through 2034.Methods COPD data by age and sex from the GBD 2019 database were analyzed for incidence,prevalence,mortality,and disability-adjusted life years(DALY)rates from 1990 to 2019.Joinpoint regression identified significant annual trends,and age-standardized incidence rates were predicted through 2034 using age-period-cohort models.Results The incidence,prevalence,mortality,and disease burden of COPD have been decreasing,and the incidence rates will continue to decrease or remain stable until 2034 in most selected countries and territories,except for a few Southeastern Asian countries.The Lao People’s Democratic Republic and Vietnam are projected to experience an increase in COPD incidence from 165.3 per 100,000 in 2019 to 177 per 100,000 in 2034 and from 179.9 per 100,000 in 2019 to 192.5 per 100,000 in 2034,respectively.Older males had a higher incidence than any other sex or age group.The sex gap in incidence rates continues to widen,though it is smaller and less significant in the younger age group than in those in the older one.Conclusion COPD rates are expected to decline until 2034 but remain a health risk,especially in countries with rising rates.Urgent action on tobacco control,air pollution,and public education is needed.展开更多
Species distribution patterns is one of the important topics in ecology and biological conservation.Although species distribution models have been intensively used in the research,the effects of spatial associations a...Species distribution patterns is one of the important topics in ecology and biological conservation.Although species distribution models have been intensively used in the research,the effects of spatial associations and spatial dependence have been rarely taken into account in the modeling processes.Recently,Joint Species Distribution Models(JSDMs)offer the opportunity to consider both environmental factors and interspecific relationships as well as the role of spatial structures.This study uses the HMSC(Hierarchical Modelling of Species Communities)framework to model the multispecies distribution of a marine fish assemblage,in which spatial associations and spatial dependence is deliberately accounted for.Three HMSC models were implemented with different structures of random effects to address the existence of spatial associations and spatial dependence,and the predictive performances at different levels of sample sizes were analyzed in the assessment.The results showed that the models with random effects could account for a larger proportion of explainable variance(32.8%),and particularly the spatial random effect model provided the best predictive performances(R_(mean)^(2)=0.31),indicating that spatial random effects could substantially influence the results of the joint species distribution.Increasing sample size had a strong effect(R_(mean)^(2)=0.24-0.31)on the predictive accuracy of the spatially-structured model than on the other models,suggesting that optimal model selection should be dependent on sample size.This study highlights the importance of incorporating spatial random effects for JSDM predictions and suggests that the choice of model structures should consider the data quality across species.展开更多
Objective:Knee osteoarthritis is one of the important causes of disability worldwide.This study aims to analyze the disease burden of knee osteoarthritis,attributable risk factors among Chinese residents from 1990 to ...Objective:Knee osteoarthritis is one of the important causes of disability worldwide.This study aims to analyze the disease burden of knee osteoarthritis,attributable risk factors among Chinese residents from 1990 to 2021,and predict the disease burden trend for 2035.Methods:Data related to knee osteoarthritis in China from 1990 to 2021,including the number of incident cases,incidence rate,number of prevalent cases,prevalence rate,and years lived with disability(YLDs),were collected from the Global Burden of Disease Study(GBD2021)database.Joinpoint regression analysis was used to assess time trends,and the Bayesian-Age-Period-Cohort(BAPC)regression model was employed for future predictions.Results:From 1990 to 2021,the number of incident cases of knee osteoarthritis among Chinese residents increased from 3.65 million to 8.51 million,a rise of 133.16%,with an average annual increase of 3.15%.The incidence rate increased from 310.33 per 100,000 to 598.31 per 100,000,a rise of 92.80%,with an average annual increase of 2.55%.The number of prevalent cases increased from 41.04 million to 110 million,a rise of 166.97%,with an average annual increase of 3.61%.The prevalence rate increased from 3488.78 per 100,000 to 7701.69 per 100,000,a rise of 120.76%,with an average annual increase of 3.00%.The number of YLDs increased from 1.34 million to 3.55 million,a rise of 165.32%,with an average annual increase of 3.59%.The YLD rate increased from 113.86 per 100,000 to 249.81 per 100,000,a rise of 119.39%,with an average annual increase of 2.99%.High BMI was the only significant attributable risk factor,with the proportion of YLDs it caused continuing to rise.Predictions for 2035:The number of incident cases is expected to decline slightly from 5.89 million in 2022 to 5.72 million in 2035.The number of prevalent cases is expected to peak at 72.42 million in 2029 and be around 72.69 million in 2035.The number of YLDs is expected to increase year by year,from 2.35 million in 2022 to 2.35 million in 2035.Conclusion:The study reveals the increasing prevalence and disease burden of knee osteoarthritis among Chinese residents,emphasizing the importance of interventions targeting controllable risk factors.Although the prediction shows a slight decline in the number of incident cases in 2035,the number of prevalent cases and years of disability are expected to remain high,indicating the need for continued monitoring and intervention.展开更多
As an effective strategy to address urban traffic congestion,traffic flow prediction has gained attention from Federated-Learning(FL)researchers due FL’s ability to preserving data privacy.However,existing methods fa...As an effective strategy to address urban traffic congestion,traffic flow prediction has gained attention from Federated-Learning(FL)researchers due FL’s ability to preserving data privacy.However,existing methods face challenges:some are too simplistic to capture complex traffic patterns effectively,and others are overly complex,leading to excessive communication overhead between cloud and edge devices.Moreover,the problem of single point failure limits their robustness and reliability in real-world applications.To tackle these challenges,this paper proposes a new method,CMBA-FL,a Communication-Mitigated and Blockchain-Assisted Federated Learning model.First,CMBA-FL improves the client model’s ability to capture temporal traffic patterns by employing the Encoder-Decoder framework for each edge device.Second,to reduce the communication overhead during federated learning,we introduce a verification method based on parameter update consistency,avoiding unnecessary parameter updates.Third,to mitigate the risk of a single point of failure,we integrate consensus mechanisms from blockchain technology.To validate the effectiveness of CMBA-FL,we assess its performance on two widely used traffic datasets.Our experimental results show that CMBA-FL reduces prediction error by 11.46%,significantly lowers communication overhead,and improves security.展开更多
Advancing the integration of artificial intelligence and polymer science requires high-quality,open-source,and large-scale datasets.However,existing polymer databases often suffer from data sparsity,lack of polymer-pr...Advancing the integration of artificial intelligence and polymer science requires high-quality,open-source,and large-scale datasets.However,existing polymer databases often suffer from data sparsity,lack of polymer-property labels,and limited accessibility,hindering system-atic modeling across property prediction tasks.Here,we present OpenPoly,a curated experimental polymer database derived from extensive lit-erature mining and manual validation,comprising 3985 unique polymer-property data points spanning 26 key properties.We further develop a multi-task benchmarking framework that evaluates property prediction using four encoding methods and eight representative models.Our re-sults highlight that the optimized degree-of-polymerization encoding coupled with Morgan fingerprints achieves an optimal trade-off between computational cost and accuracy.In data-scarce condition,XGBoost outperforms deep learning models on key properties such as dielectric con-stant,glass transition temperature,melting point,and mechanical strength,achieving R2 scores of 0.65-0.87.To further showcase the practical utility of the database,we propose potential polymers for two energy-relevant applications:high temperature polymer dielectrics and fuel cell membranes.By offering a consistent and accessible benchmark and database,OpenPoly paves the way for more accurate polymer-property modeling and fosters data-driven advances in polymer genome engineering.展开更多
Rotor blade is one of the most significant components of helicopters. But due to its highspeed rotation characteristics, it is difficult to collect the vibration signals during the flight stage.Moreover, sensors are h...Rotor blade is one of the most significant components of helicopters. But due to its highspeed rotation characteristics, it is difficult to collect the vibration signals during the flight stage.Moreover, sensors are highly susceptible to damage resulting in the failure of the measurement.In order to make signal predictions for the damaged sensors, an operational modal analysis(OMA) together with the virtual sensing(VS) technology is proposed in this paper. This paper discusses two situations, i.e., mode shapes measured by all sensors(both normal and damaged) can be obtained using OMA, and mode shapes measured by some sensors(only including normal) can be obtained using OMA. For the second situation, it is necessary to use finite element(FE) analysis to supplement the missing mode shapes of damaged sensor. In order to improve the correlation between the FE model and the real structure, the FE mode shapes are corrected using the local correspondence(LC) principle and mode shapes measured by some sensors(only including normal).Then, based on the VS technology, the vibration signals of the damaged sensors during the flight stage can be accurately predicted using the identified mode shapes(obtained based on OMA and FE analysis) and the normal sensors signals. Given the high degrees of freedom(DOFs) in the FE mode shapes, this approach can also be used to predict vibration data at locations without sensors. The effectiveness and robustness of the proposed method is verified through finite element simulation, experiment as well as the actual flight test. The present work can be further used in the fault diagnosis and damage identification for rotor blade of helicopters.展开更多
Steels are widely used as structural materials,making them essential for supporting our lives and industries.However,further improving the comprehensive properties of steel through traditional trial-and-error methods ...Steels are widely used as structural materials,making them essential for supporting our lives and industries.However,further improving the comprehensive properties of steel through traditional trial-and-error methods becomes challenging due to the continuous development and numerous processing parameters involved in steel production.To address this challenge,the application of machine learning methods becomes crucial in establishing complex relationships between manufacturing processes and steel performance.This review begins with a general overview of machine learning methods and subsequently introduces various performance predictions in steel materials.The classification of performance pre-diction was used to assess the current application of machine learning model-assisted design.Several important issues,such as data source and characteristics,intermediate features,algorithm optimization,key feature analysis,and the role of environmental factors,were summarized and analyzed.These insights will be beneficial and enlightening to future research endeavors in this field.展开更多
In recent years,deep learning methods have gradually been applied to prediction tasks related to Arctic sea ice concentration,but relatively little research has been conducted for larger spatial and temporal scales,ma...In recent years,deep learning methods have gradually been applied to prediction tasks related to Arctic sea ice concentration,but relatively little research has been conducted for larger spatial and temporal scales,mainly due to the limited time coverage of observations and reanalysis data.Meanwhile,deep learning predictions of sea ice thickness(SIT)have yet to receive ample attention.In this study,two data-driven deep learning(DL)models are built based on the ConvLSTM and fully convolutional U-net(FC-Unet)algorithms and trained using CMIP6 historical simulations for transfer learning and fine-tuned using reanalysis/observations.These models enable monthly predictions of Arctic SIT without considering the complex physical processes involved.Through comprehensive assessments of prediction skills by season and region,the results suggest that using a broader set of CMIP6 data for transfer learning,as well as incorporating multiple climate variables as predictors,contribute to better prediction results,although both DL models can effectively predict the spatiotemporal features of SIT anomalies.Regarding the predicted SIT anomalies of the FC-Unet model,the spatial correlations with reanalysis reach an average level of 89%over all months,while the temporal anomaly correlation coefficients are close to unity in most cases.The models also demonstrate robust performances in predicting SIT and SIE during extreme events.The effectiveness and reliability of the proposed deep transfer learning models in predicting Arctic SIT can facilitate more accurate pan-Arctic predictions,aiding climate change research and real-time business applications.展开更多
To compare finite element analysis(FEA)predictions and stereovision digital image correlation(StereoDIC)strain measurements at the same spatial positions throughout a region of interest,a field comparison procedure is...To compare finite element analysis(FEA)predictions and stereovision digital image correlation(StereoDIC)strain measurements at the same spatial positions throughout a region of interest,a field comparison procedure is developed.The procedure includes(a)conversion of the finite element data into a triangular mesh,(b)selection of a common coordinate system,(c)determination of the rigid body transformation to place both measurements and FEA data in the same system and(d)interpolation of the FEA nodal information to the same spatial locations as the StereoDIC measurements using barycentric coordinates.For an aluminum Al-6061 double edge notched tensile specimen,FEA results are obtained using both the von Mises isotropic yield criterion and Hill’s quadratic anisotropic yield criterion,with the unknown Hill model parameters determined using full-field specimen strain measurements for the nominally plane stress specimen.Using Hill’s quadratic anisotropic yield criterion,the point-by-point comparison of experimentally based full-field strains and stresses to finite element predictions are shown to be in excellent agreement,confirming the effectiveness of the field comparison process.展开更多
Predictive Business Process Monitoring(PBPM)is a significant research area in Business Process Management(BPM)aimed at accurately forecasting future behavioral events.At present,deep learning methods are widely cited ...Predictive Business Process Monitoring(PBPM)is a significant research area in Business Process Management(BPM)aimed at accurately forecasting future behavioral events.At present,deep learning methods are widely cited in PBPM research,but no method has been effective in fusing data information into the control flow for multi-perspective process prediction.Therefore,this paper proposes a process prediction method based on the hierarchical BERT and multi-perspective data fusion.Firstly,the first layer BERT network learns the correlations between different category attribute data.Then,the attribute data is integrated into a weighted event-level feature vector and input into the second layer BERT network to learn the impact and priority relationship of each event on future predicted events.Next,the multi-head attention mechanism within the framework is visualized for analysis,helping to understand the decision-making logic of the framework and providing visual predictions.Finally,experimental results show that the predictive accuracy of the framework surpasses the current state-of-the-art research methods and significantly enhances the predictive performance of BPM.展开更多
Understanding the nonlinear relationship between hydrological response and key factors and the cause behind this relationship is vital for water resource management and earth system model.In this study,we undertook se...Understanding the nonlinear relationship between hydrological response and key factors and the cause behind this relationship is vital for water resource management and earth system model.In this study,we undertook several steps to explore the relationship.Initially,we partitioned runoff response change(RRC)into multiple components associated with climate and catchment properties,and examined the spatial patterns and smoothness indicated by the Moran's Index of RRC across the contiguous United States(CONUS).Subsequently,we employed a machine learning model to predict RRC using catchment attribute predictors encompassing climate,topography,hydrology,soil,land use/cover,and geology.Additionally,we identified the primary factors influencing RRC and quantified how these key factors control RRC by employing the accumulated local effect,which allows for the representation of not only dominant but also secondary effects.Finally,we explored the relationship between ecoregion patterns,climate gradients,and the distribution of RRC across CONUS.Our findings indicate that:(1)RRC demonstrating significant connections between catchments tends to be well predicted by catchment attributes in space;(2)climate,hydrology,and topography emerge as the top three key attributes nonlinearly influencing the RRC patterns,with their second-order effects determining the heterogeneous patterns of RRC;and(3)local Moran's I signifies a collaborative relationship between the patterns of RRC and their spatial smoothness,climate space,and ecoregions.展开更多
Background:This study aims to explore the involvement of ferroptosis-related genes and pathogenesis in pancreatic cancer and predict potential therapeutic interventions using Traditional Chinese Medicine(TCM).Methods:...Background:This study aims to explore the involvement of ferroptosis-related genes and pathogenesis in pancreatic cancer and predict potential therapeutic interventions using Traditional Chinese Medicine(TCM).Methods:We utilized gene expression datasets,ferroptosis upregulated genes and applied machine learning algorithms,including LASSO and SVM-RFE,to identify key ferroptosis-related genes in pancreatic cancer.Perform Gene Ontology,Kyoto Encyclopedia of Genes and Genomes,and Disease Ontology enrichment analysis,immune infiltration analysis and correlation analysis between immune infiltrating cells and characteristic genes on differentially expressed genes using the R software package.Retrieve potential traditional Chinese medicine for targeted ferroptosis gene therapy for pancreatic cancer through Coremine and Herb databases.Results:Seventeen feature genes were identified,with significant implications for immune cell infiltration in pancreatic cancer.The results of immune cell infiltration analysis showed that B cells naive,B cells memory,T cells regulatory,and M0 macrophages were significantly upregulated in pancreatic cancer patients;Mast cells resting were significantly downregulated.Chinese herbal medicines such as ginkgo,turmeric,ginseng,Codonopsis pilosula,Zedoary turmeric,deer tendons,senna leaves,Guanmu Tong,Huangqi,and Banzhilian are potential drugs for targeted ferroptosis gene therapy for pancreatic cancer.Conclusion:TIMP1 emerged as a key gene,with several TCM herbs predicted to modulate its expression,offering new avenues for treatment.展开更多
One of the significant issues the insurance industry faces is its ability to predict future claims related to individual policyholders.As risk varies from one policyholder to another,the industry has faced the challen...One of the significant issues the insurance industry faces is its ability to predict future claims related to individual policyholders.As risk varies from one policyholder to another,the industry has faced the challenge of using various risk factors to accurately predict the likelihood of claims by policyholders using historical data.Traditional machine-learning models that use neural networks are recognized as exceptional algorithms with predictive capabilities.This study aims to develop a deep learning model using sequential deep regression techniques for insurance claim prediction using historical data obtained from Kaggle with 1339 cases and eight variables.This study adopted a sequential model in Keras and compared the model with ReLU and Swish as activation functions.The performance metrics used during the training to evaluate the model performance are R2 score,mean square error and mean percentage error with values of 0.5%,1.17%and 23.5%,respectively obtained using ReLU,while 0.7%,0.82%,and 21.3%,obtained using Swish function.Although the results of both models using ReLU and Swish were fairly tolerable,the performance metrics obtained of the model using Swish,as the activation function tends to perform more satisfactorily than that of ReLU.However,to investigate the model performance deeply,therefore,this study recommends that more interest be channeled to the interpretability and explainability of the proposed model and the provisions of AI technologies by insurance industries to enhance accurate claim prediction and minimize losses.展开更多
基金supported by a major project of the Zhejiang Natural Science Foundation(LD21G030001).
文摘Objective The Asia-Pacific region has a high chronic obstructive pulmonary disease(COPD)burden,but studies on its trends are limited.Using the Global Burden of Disease(GBD)2019 data,we analyzed COPD trends in 36 countries and territories from 1990 to 2019 and predicted future incidence trends through 2034.Methods COPD data by age and sex from the GBD 2019 database were analyzed for incidence,prevalence,mortality,and disability-adjusted life years(DALY)rates from 1990 to 2019.Joinpoint regression identified significant annual trends,and age-standardized incidence rates were predicted through 2034 using age-period-cohort models.Results The incidence,prevalence,mortality,and disease burden of COPD have been decreasing,and the incidence rates will continue to decrease or remain stable until 2034 in most selected countries and territories,except for a few Southeastern Asian countries.The Lao People’s Democratic Republic and Vietnam are projected to experience an increase in COPD incidence from 165.3 per 100,000 in 2019 to 177 per 100,000 in 2034 and from 179.9 per 100,000 in 2019 to 192.5 per 100,000 in 2034,respectively.Older males had a higher incidence than any other sex or age group.The sex gap in incidence rates continues to widen,though it is smaller and less significant in the younger age group than in those in the older one.Conclusion COPD rates are expected to decline until 2034 but remain a health risk,especially in countries with rising rates.Urgent action on tobacco control,air pollution,and public education is needed.
基金supported by the National Key R&D Program of China(No.2022YFD2401301)。
文摘Species distribution patterns is one of the important topics in ecology and biological conservation.Although species distribution models have been intensively used in the research,the effects of spatial associations and spatial dependence have been rarely taken into account in the modeling processes.Recently,Joint Species Distribution Models(JSDMs)offer the opportunity to consider both environmental factors and interspecific relationships as well as the role of spatial structures.This study uses the HMSC(Hierarchical Modelling of Species Communities)framework to model the multispecies distribution of a marine fish assemblage,in which spatial associations and spatial dependence is deliberately accounted for.Three HMSC models were implemented with different structures of random effects to address the existence of spatial associations and spatial dependence,and the predictive performances at different levels of sample sizes were analyzed in the assessment.The results showed that the models with random effects could account for a larger proportion of explainable variance(32.8%),and particularly the spatial random effect model provided the best predictive performances(R_(mean)^(2)=0.31),indicating that spatial random effects could substantially influence the results of the joint species distribution.Increasing sample size had a strong effect(R_(mean)^(2)=0.24-0.31)on the predictive accuracy of the spatially-structured model than on the other models,suggesting that optimal model selection should be dependent on sample size.This study highlights the importance of incorporating spatial random effects for JSDM predictions and suggests that the choice of model structures should consider the data quality across species.
文摘Objective:Knee osteoarthritis is one of the important causes of disability worldwide.This study aims to analyze the disease burden of knee osteoarthritis,attributable risk factors among Chinese residents from 1990 to 2021,and predict the disease burden trend for 2035.Methods:Data related to knee osteoarthritis in China from 1990 to 2021,including the number of incident cases,incidence rate,number of prevalent cases,prevalence rate,and years lived with disability(YLDs),were collected from the Global Burden of Disease Study(GBD2021)database.Joinpoint regression analysis was used to assess time trends,and the Bayesian-Age-Period-Cohort(BAPC)regression model was employed for future predictions.Results:From 1990 to 2021,the number of incident cases of knee osteoarthritis among Chinese residents increased from 3.65 million to 8.51 million,a rise of 133.16%,with an average annual increase of 3.15%.The incidence rate increased from 310.33 per 100,000 to 598.31 per 100,000,a rise of 92.80%,with an average annual increase of 2.55%.The number of prevalent cases increased from 41.04 million to 110 million,a rise of 166.97%,with an average annual increase of 3.61%.The prevalence rate increased from 3488.78 per 100,000 to 7701.69 per 100,000,a rise of 120.76%,with an average annual increase of 3.00%.The number of YLDs increased from 1.34 million to 3.55 million,a rise of 165.32%,with an average annual increase of 3.59%.The YLD rate increased from 113.86 per 100,000 to 249.81 per 100,000,a rise of 119.39%,with an average annual increase of 2.99%.High BMI was the only significant attributable risk factor,with the proportion of YLDs it caused continuing to rise.Predictions for 2035:The number of incident cases is expected to decline slightly from 5.89 million in 2022 to 5.72 million in 2035.The number of prevalent cases is expected to peak at 72.42 million in 2029 and be around 72.69 million in 2035.The number of YLDs is expected to increase year by year,from 2.35 million in 2022 to 2.35 million in 2035.Conclusion:The study reveals the increasing prevalence and disease burden of knee osteoarthritis among Chinese residents,emphasizing the importance of interventions targeting controllable risk factors.Although the prediction shows a slight decline in the number of incident cases in 2035,the number of prevalent cases and years of disability are expected to remain high,indicating the need for continued monitoring and intervention.
基金supported by the National Natural Science Foundation of China under Grant No.U20A20182.
文摘As an effective strategy to address urban traffic congestion,traffic flow prediction has gained attention from Federated-Learning(FL)researchers due FL’s ability to preserving data privacy.However,existing methods face challenges:some are too simplistic to capture complex traffic patterns effectively,and others are overly complex,leading to excessive communication overhead between cloud and edge devices.Moreover,the problem of single point failure limits their robustness and reliability in real-world applications.To tackle these challenges,this paper proposes a new method,CMBA-FL,a Communication-Mitigated and Blockchain-Assisted Federated Learning model.First,CMBA-FL improves the client model’s ability to capture temporal traffic patterns by employing the Encoder-Decoder framework for each edge device.Second,to reduce the communication overhead during federated learning,we introduce a verification method based on parameter update consistency,avoiding unnecessary parameter updates.Third,to mitigate the risk of a single point of failure,we integrate consensus mechanisms from blockchain technology.To validate the effectiveness of CMBA-FL,we assess its performance on two widely used traffic datasets.Our experimental results show that CMBA-FL reduces prediction error by 11.46%,significantly lowers communication overhead,and improves security.
基金financially supported by the National Natural Science Foundation of China (Nos. 92372126,52373203)the Excellent Young Scientists Fund Program
文摘Advancing the integration of artificial intelligence and polymer science requires high-quality,open-source,and large-scale datasets.However,existing polymer databases often suffer from data sparsity,lack of polymer-property labels,and limited accessibility,hindering system-atic modeling across property prediction tasks.Here,we present OpenPoly,a curated experimental polymer database derived from extensive lit-erature mining and manual validation,comprising 3985 unique polymer-property data points spanning 26 key properties.We further develop a multi-task benchmarking framework that evaluates property prediction using four encoding methods and eight representative models.Our re-sults highlight that the optimized degree-of-polymerization encoding coupled with Morgan fingerprints achieves an optimal trade-off between computational cost and accuracy.In data-scarce condition,XGBoost outperforms deep learning models on key properties such as dielectric con-stant,glass transition temperature,melting point,and mechanical strength,achieving R2 scores of 0.65-0.87.To further showcase the practical utility of the database,we propose potential polymers for two energy-relevant applications:high temperature polymer dielectrics and fuel cell membranes.By offering a consistent and accessible benchmark and database,OpenPoly paves the way for more accurate polymer-property modeling and fosters data-driven advances in polymer genome engineering.
基金supported by grants from the High-Level Oversea Talent Introduction Plan,Chinathe Special Fund for Basic Scientific Research in Central Universities of China-Doctoral Research and Innovation Fund Project,China(No.3072023CFJ0206).
文摘Rotor blade is one of the most significant components of helicopters. But due to its highspeed rotation characteristics, it is difficult to collect the vibration signals during the flight stage.Moreover, sensors are highly susceptible to damage resulting in the failure of the measurement.In order to make signal predictions for the damaged sensors, an operational modal analysis(OMA) together with the virtual sensing(VS) technology is proposed in this paper. This paper discusses two situations, i.e., mode shapes measured by all sensors(both normal and damaged) can be obtained using OMA, and mode shapes measured by some sensors(only including normal) can be obtained using OMA. For the second situation, it is necessary to use finite element(FE) analysis to supplement the missing mode shapes of damaged sensor. In order to improve the correlation between the FE model and the real structure, the FE mode shapes are corrected using the local correspondence(LC) principle and mode shapes measured by some sensors(only including normal).Then, based on the VS technology, the vibration signals of the damaged sensors during the flight stage can be accurately predicted using the identified mode shapes(obtained based on OMA and FE analysis) and the normal sensors signals. Given the high degrees of freedom(DOFs) in the FE mode shapes, this approach can also be used to predict vibration data at locations without sensors. The effectiveness and robustness of the proposed method is verified through finite element simulation, experiment as well as the actual flight test. The present work can be further used in the fault diagnosis and damage identification for rotor blade of helicopters.
基金supported by the National Natural Science Foundation of China (No.51701061)the Natural Science Foundation of Hebei Province (Nos.E2023202047 and E2021202075)+1 种基金the Key-Area R&D Program of Guangdong Province (No.2020B0101340004)Guangdong Academy of Science (2021GDASYL-20210102002).
文摘Steels are widely used as structural materials,making them essential for supporting our lives and industries.However,further improving the comprehensive properties of steel through traditional trial-and-error methods becomes challenging due to the continuous development and numerous processing parameters involved in steel production.To address this challenge,the application of machine learning methods becomes crucial in establishing complex relationships between manufacturing processes and steel performance.This review begins with a general overview of machine learning methods and subsequently introduces various performance predictions in steel materials.The classification of performance pre-diction was used to assess the current application of machine learning model-assisted design.Several important issues,such as data source and characteristics,intermediate features,algorithm optimization,key feature analysis,and the role of environmental factors,were summarized and analyzed.These insights will be beneficial and enlightening to future research endeavors in this field.
基金supported by the National Natural Science Foundation of China(Grant Nos.41976193 and 42176243).
文摘In recent years,deep learning methods have gradually been applied to prediction tasks related to Arctic sea ice concentration,but relatively little research has been conducted for larger spatial and temporal scales,mainly due to the limited time coverage of observations and reanalysis data.Meanwhile,deep learning predictions of sea ice thickness(SIT)have yet to receive ample attention.In this study,two data-driven deep learning(DL)models are built based on the ConvLSTM and fully convolutional U-net(FC-Unet)algorithms and trained using CMIP6 historical simulations for transfer learning and fine-tuned using reanalysis/observations.These models enable monthly predictions of Arctic SIT without considering the complex physical processes involved.Through comprehensive assessments of prediction skills by season and region,the results suggest that using a broader set of CMIP6 data for transfer learning,as well as incorporating multiple climate variables as predictors,contribute to better prediction results,although both DL models can effectively predict the spatiotemporal features of SIT anomalies.Regarding the predicted SIT anomalies of the FC-Unet model,the spatial correlations with reanalysis reach an average level of 89%over all months,while the temporal anomaly correlation coefficients are close to unity in most cases.The models also demonstrate robust performances in predicting SIT and SIE during extreme events.The effectiveness and reliability of the proposed deep transfer learning models in predicting Arctic SIT can facilitate more accurate pan-Arctic predictions,aiding climate change research and real-time business applications.
基金Financial support provided by Correlated Solutions Incorporated to perform StereoDIC experimentsthe Department of Mechanical Engineering at the University of South Carolina for simulation studies is deeply appreciated.
文摘To compare finite element analysis(FEA)predictions and stereovision digital image correlation(StereoDIC)strain measurements at the same spatial positions throughout a region of interest,a field comparison procedure is developed.The procedure includes(a)conversion of the finite element data into a triangular mesh,(b)selection of a common coordinate system,(c)determination of the rigid body transformation to place both measurements and FEA data in the same system and(d)interpolation of the FEA nodal information to the same spatial locations as the StereoDIC measurements using barycentric coordinates.For an aluminum Al-6061 double edge notched tensile specimen,FEA results are obtained using both the von Mises isotropic yield criterion and Hill’s quadratic anisotropic yield criterion,with the unknown Hill model parameters determined using full-field specimen strain measurements for the nominally plane stress specimen.Using Hill’s quadratic anisotropic yield criterion,the point-by-point comparison of experimentally based full-field strains and stresses to finite element predictions are shown to be in excellent agreement,confirming the effectiveness of the field comparison process.
基金Supported by the National Natural Science Foundation,China(No.61402011)the Open Project Program of the Key Laboratory of Embedded System and Service Computing of Ministry of Education(No.ESSCKF2021-05).
文摘Predictive Business Process Monitoring(PBPM)is a significant research area in Business Process Management(BPM)aimed at accurately forecasting future behavioral events.At present,deep learning methods are widely cited in PBPM research,but no method has been effective in fusing data information into the control flow for multi-perspective process prediction.Therefore,this paper proposes a process prediction method based on the hierarchical BERT and multi-perspective data fusion.Firstly,the first layer BERT network learns the correlations between different category attribute data.Then,the attribute data is integrated into a weighted event-level feature vector and input into the second layer BERT network to learn the impact and priority relationship of each event on future predicted events.Next,the multi-head attention mechanism within the framework is visualized for analysis,helping to understand the decision-making logic of the framework and providing visual predictions.Finally,experimental results show that the predictive accuracy of the framework surpasses the current state-of-the-art research methods and significantly enhances the predictive performance of BPM.
基金National Natural Science Foundation of China,No.U2243203,No.51979069Natural Science Foundation of Jiangsu Province,China,No.BK20211202Research Council of Norway,No.FRINATEK Project 274310。
文摘Understanding the nonlinear relationship between hydrological response and key factors and the cause behind this relationship is vital for water resource management and earth system model.In this study,we undertook several steps to explore the relationship.Initially,we partitioned runoff response change(RRC)into multiple components associated with climate and catchment properties,and examined the spatial patterns and smoothness indicated by the Moran's Index of RRC across the contiguous United States(CONUS).Subsequently,we employed a machine learning model to predict RRC using catchment attribute predictors encompassing climate,topography,hydrology,soil,land use/cover,and geology.Additionally,we identified the primary factors influencing RRC and quantified how these key factors control RRC by employing the accumulated local effect,which allows for the representation of not only dominant but also secondary effects.Finally,we explored the relationship between ecoregion patterns,climate gradients,and the distribution of RRC across CONUS.Our findings indicate that:(1)RRC demonstrating significant connections between catchments tends to be well predicted by catchment attributes in space;(2)climate,hydrology,and topography emerge as the top three key attributes nonlinearly influencing the RRC patterns,with their second-order effects determining the heterogeneous patterns of RRC;and(3)local Moran's I signifies a collaborative relationship between the patterns of RRC and their spatial smoothness,climate space,and ecoregions.
基金supported by the Modern Traditional Chinese Medicine Haihe Laboratory science and technology project(22HHZYSS00005)and the National Administration of Traditional Chinese Medicine Young Qihuang Scholar Project.
文摘Background:This study aims to explore the involvement of ferroptosis-related genes and pathogenesis in pancreatic cancer and predict potential therapeutic interventions using Traditional Chinese Medicine(TCM).Methods:We utilized gene expression datasets,ferroptosis upregulated genes and applied machine learning algorithms,including LASSO and SVM-RFE,to identify key ferroptosis-related genes in pancreatic cancer.Perform Gene Ontology,Kyoto Encyclopedia of Genes and Genomes,and Disease Ontology enrichment analysis,immune infiltration analysis and correlation analysis between immune infiltrating cells and characteristic genes on differentially expressed genes using the R software package.Retrieve potential traditional Chinese medicine for targeted ferroptosis gene therapy for pancreatic cancer through Coremine and Herb databases.Results:Seventeen feature genes were identified,with significant implications for immune cell infiltration in pancreatic cancer.The results of immune cell infiltration analysis showed that B cells naive,B cells memory,T cells regulatory,and M0 macrophages were significantly upregulated in pancreatic cancer patients;Mast cells resting were significantly downregulated.Chinese herbal medicines such as ginkgo,turmeric,ginseng,Codonopsis pilosula,Zedoary turmeric,deer tendons,senna leaves,Guanmu Tong,Huangqi,and Banzhilian are potential drugs for targeted ferroptosis gene therapy for pancreatic cancer.Conclusion:TIMP1 emerged as a key gene,with several TCM herbs predicted to modulate its expression,offering new avenues for treatment.
文摘One of the significant issues the insurance industry faces is its ability to predict future claims related to individual policyholders.As risk varies from one policyholder to another,the industry has faced the challenge of using various risk factors to accurately predict the likelihood of claims by policyholders using historical data.Traditional machine-learning models that use neural networks are recognized as exceptional algorithms with predictive capabilities.This study aims to develop a deep learning model using sequential deep regression techniques for insurance claim prediction using historical data obtained from Kaggle with 1339 cases and eight variables.This study adopted a sequential model in Keras and compared the model with ReLU and Swish as activation functions.The performance metrics used during the training to evaluate the model performance are R2 score,mean square error and mean percentage error with values of 0.5%,1.17%and 23.5%,respectively obtained using ReLU,while 0.7%,0.82%,and 21.3%,obtained using Swish function.Although the results of both models using ReLU and Swish were fairly tolerable,the performance metrics obtained of the model using Swish,as the activation function tends to perform more satisfactorily than that of ReLU.However,to investigate the model performance deeply,therefore,this study recommends that more interest be channeled to the interpretability and explainability of the proposed model and the provisions of AI technologies by insurance industries to enhance accurate claim prediction and minimize losses.
