Accurate prediction of pure component physiochemical properties is crucial for process integration, multiscale modelling, and optimization. In this work, an enhanced framework for pure component property prediction by...Accurate prediction of pure component physiochemical properties is crucial for process integration, multiscale modelling, and optimization. In this work, an enhanced framework for pure component property prediction by using explainable machine learning methods is proposed. In this framework, the molecular representation method based on the connectivity matrix effectively considers atomic bonding relationships to automatically generate features. The supervised machine learning model random forest is applied for feature ranking and pooling. The adjusted R^(2) is introduced to penalize the inclusion of additional features, providing an assessment of the true contribution of features. The prediction results for normal boiling point (T_(b)), liquid molar volume (L_(mv)), critical temperature (T_(c)) and critical pressure (P_(c)) obtained using Artificial Neural Network and Gaussian Process Regression models confirm the accuracy of the molecular representation method. Comparison with GC based models shows that the root-mean-square error on the test set can be reduced by up to 83.8%. To enhance the interpretability of the model, a feature analysis method based on Shapley values is employed to determine the contribution of each feature to the property predictions. The results indicate that using the feature pooling method reduces the number of features from 13316 to 100 without compromising model accuracy. The feature analysis results for Tb, Lmv, Tc, and Pc confirms that different molecular properties are influenced by different structural features, aligning with mechanistic interpretations. In conclusion, the proposed framework is demonstrated to be feasible and provides a solid foundation for mixture component reconstruction and process integration modelling.展开更多
Glacier mass balance is a key indicator of glacier health and climate change sensitivity.Influencing factors include both climatic and nonclimatic elements,forming a complex set of drivers.There is a lack of quantitat...Glacier mass balance is a key indicator of glacier health and climate change sensitivity.Influencing factors include both climatic and nonclimatic elements,forming a complex set of drivers.There is a lack of quantitative analysis of these composite factors,particularly in climate-typical regions like the Tanggula Mountains on the central Tibetan Plateau.We collected data on various factors affecting glacier mass balance from 2000 to 2020,including climate variables,topographic variables,geometric parameters,and glacier dynamics.We utilized linear regression models,ensemble learning models,and Open Global Glacier Model(OGGM)to analyze glacier mass balance changes in the Tanggula Mountains.Results indicate that linear models explain 58%of the variance in glacier mass balance,with seasonal temperature and precipitation having significant impacts.Our findings show that ensemble learning models made the explanations 5.2%more accurate by including the impact of topographic and geometric factors such as the average glacier height,the slope of the glacier tongue,the speed of the ice flow,and the area of the glacier.Interpretable machine learning identified the spatial distribution of positive and negative impacts of these characteristics and the interaction between glacier topography and ice dynamics.Finally,we predicted the responses of glaciers of different sizes to future climate change based on the results of interpretable machine learning.It was found that relatively large glaciers(>1 km~2)are likely to persist until the end of this century under low emission scenarios,whereas small glaciers(<1 km~2)are expected to nearly disappear by 2080 under any emission scenario.Our research provides technical support for improving glacier change modeling and protection on the Tibetan Plateau.展开更多
Background:Frailty in older adults is linked to increased risks and lower quality of life.Pre-frailty,a condition preceding frailty,is intervenable,but its determinants and assessment are challenging.This study aims t...Background:Frailty in older adults is linked to increased risks and lower quality of life.Pre-frailty,a condition preceding frailty,is intervenable,but its determinants and assessment are challenging.This study aims to develop and validate an explainable machine learning model for pre-frailty risk assessment among community-dwelling older adults.Methods:The study included 3141 adults aged 60 or above from the China Health and Retirement Longitudinal Study.Pre-frailty was characterized by one or two criteria from the physical frailty phenotype scale.We extracted 80 distinct features across seven dimensions to evaluate pre-frailty risk.A model was constructed using recursive feature elimination and a stacking-CatBoost distillation module on 80%of the sample and validated on a separate 20%holdout data set.Results:The study used data from 2508 community-dwelling older adults(mean age,67.24 years[range,60–96];1215[48.44%]females)to develop a pre-frailty risk assessment model.We selected 57 predictive features and built a distilled CatBoost model,which achieved the highest discrimination(AUROC:0.7560[95%CI:0.7169,0.7928])on the 20%holdout data set.The living city,BMI,and peak expiratory flow(PEF)were the three most significant contributors to pre-frailty risk.Physical and environmental factors were the top 2 impactful feature dimensions.Conclusions:An accurate and interpretable pre-frailty risk assessment framework using state-of-the-art machine learning techniques and explanation methods has been developed.Our framework incorporates a wide range of features and determinants,allowing for a comprehensive and nuanced understanding of pre-frailty risk.展开更多
Nowadays,quantum machine learning is attracting great interest in a wide range offields due to its potential superior performance and capabilities.The massive increase in computational capacity and speed of quantum com...Nowadays,quantum machine learning is attracting great interest in a wide range offields due to its potential superior performance and capabilities.The massive increase in computational capacity and speed of quantum computers can lead to a quantum leap in the healthcarefield.Heart disease seriously threa-tens human health since it is the leading cause of death worldwide.Quantum machine learning methods can propose effective solutions to predict heart disease and aid in early diagnosis.In this study,an ensemble machine learning model based on quantum machine learning classifiers is proposed to predict the risk of heart disease.The proposed model is a bagging ensemble learning model where a quantum support vector classifier was used as a base classifier.Further-more,in order to make the model’s outcomes more explainable,the importance of every single feature in the prediction is computed and visualized using SHapley Additive exPlanations(SHAP)framework.In the experimental study,other stand-alone quantum classifiers,namely,Quantum Support Vector Classifier(QSVC),Quantum Neural Network(QNN),and Variational Quantum Classifier(VQC)are applied and compared with classical machine learning classifiers such as Sup-port Vector Machine(SVM),and Artificial Neural Network(ANN).The experi-mental results on the Cleveland dataset reveal the superiority of QSVC compared to the others,which explains its use in the proposed bagging model.The Bagging-QSVC model outperforms all aforementioned classifiers with an accuracy of 90.16%while showing great competitiveness compared to some state-of-the-art models using the same dataset.The results of the study indicate that quantum machine learning classifiers perform better than classical machine learning classi-fiers in predicting heart disease.In addition,the study reveals that the bagging ensemble learning technique is effective in improving the prediction accuracy of quantum classifiers.展开更多
Retaining walls are utilized to support the earth and prevent the soil from spreading with natural slope angles where there are differences in the elevation of ground surfaces.As the need for retaining structures incr...Retaining walls are utilized to support the earth and prevent the soil from spreading with natural slope angles where there are differences in the elevation of ground surfaces.As the need for retaining structures increases,the use of retaining walls is increasing.The retaining walls,which increase the stability of levels,are economical and meet existing adverse conditions.A considerable amount of retaining walls is made from steel-reinforced concrete.The construction of reinforced concrete retaining walls can be costly due to its components.For this reason,the optimum cost should be targeted in the design of retaining walls.This study presents an artificial neural network(ANN)model developed to predict the optimum dimensions of a retaining wall using soil properties,material properties,and external loading conditions.The dataset utilized to train the ANN model is generated with the Flower Pollination Algorithm.The target variables in the dataset are the length of the heel(y1),length of the toe(y2),thickness of the stem(top)(y3),thickness of the stem(bottom)(y4),foundation base thickness(y5)and cost(y6)and these are estimated by utilizing an ANN model based on the height of the wall(x1),material unit weight(x2),wall friction angle(x3),surcharge load(x4),concrete cost per m3(x5),steel cost per ton(x6)and the soil class(x7).The model is formulated and trained as a multi-output regression model,as all outputs are numeric and continuous.The training and evaluation of the model results in a high prediction performance(R20.99).In addition,the impacts of different input features on the model>predictions are revealed using the SHapley Additive exPlanations(SHAP)algorithm.The study demonstrates that when trained with a large dataset,ANN models perform very well by predicting the optimal cost with high performance.展开更多
As more medical data become digitalized,machine learning is regarded as a promising tool for constructing medical decision support systems.Even with vast medical data volumes,machine learning is still not fully exploi...As more medical data become digitalized,machine learning is regarded as a promising tool for constructing medical decision support systems.Even with vast medical data volumes,machine learning is still not fully exploiting its potential because the data usually sits in data silos,and privacy and security regulations restrict their access and use.To address these issues,we built a secured and explainable machine learning framework,called explainable federated XGBoost(EXPERTS),which can share valuable information among different medical institutions to improve the learning results without sharing the patients’ data.It also reveals how the machine makes a decision through eigenvalues to offer a more insightful answer to medical professionals.To study the performance,we evaluate our approach by real-world datasets,and our approach outperforms the benchmark algorithms under both federated learning and non-federated learning frameworks.展开更多
This study develops an approach consisting of a stacking model integrated with a multi-objective optimisation algorithm aimed at predicting and optimising the ecological performance of buildings.The integrated model c...This study develops an approach consisting of a stacking model integrated with a multi-objective optimisation algorithm aimed at predicting and optimising the ecological performance of buildings.The integrated model consists of five base models and a meta-model,which significantly improves the prediction performance.Specifically,the R2 value was improved by 9.19% and the error metrics MAE,MSE,MAPE,and CVRMSE were reduced by 69.47%,79.88%,67.32%,and 57.02%,respectively,compared to the single prediction model.According to the research on interpretable machine learning,adding the SHAP value gives us a deeper understanding of the impact of each architectural design parameter on the performance.In the multi-objective optimisation part,we used the NSGA-Ⅲ algorithm to successfully improve the energy efficiency,daylight utilisation and thermal comfort of the building.Specifically,the optimal design solution reduces the energy use intensity by 31.6 kWh/m^(2),improves the useful daylight index by 39%,and modulated the thermal comfort index,resulting in a decrement of 0.69℃ for the summer season and an enhancement of 0.64℃ for the winter season,respectively.Overall,this study provides building designers and decision makers with a tool to make better design decisions at an early stage to achieve a better combination of energy efficiency,daylight utilisation and thermal comfort optimisation in an integrated manner,providing an important support for achieving sustainable building design.展开更多
Participants in oil markets are increasingly aware of the climate risks posed by frequent extreme weather.This paper examines the role of extremely high-temperature weather information in predicting oil futures prices...Participants in oil markets are increasingly aware of the climate risks posed by frequent extreme weather.This paper examines the role of extremely high-temperature weather information in predicting oil futures prices on the China International Energy Exchange(INE).An extreme high-temperature weather index(HTI)is developed on the basis of meteorological data at INE’s crude oil production and storage sites.The local interpretable model-agnostic explanations(LIME)and accumulated local effects(ALE)methods are used to compare the predictive contribution of the HTI with that of 15 common predictors.The results indicate that the HTI enhances the out-of-sample accuracy of five classical prediction models for INE oil prices.The recurrent neural network(RNN)model exhibits superior out-of-sample forecast performance,with an MAE of 14.379,an RMSE of 19.624,and a DS of 66.67%.The predictive importance of the HTI in the best RNN model ranks third in most test instances,surpassing conventional oil price predictors such as stock market indicators.The ALE analysis reveals a positive correlation between extremely high-temperature weather and INE oil prices.These findings can help investors and oil market regulators improve oil price forecast accuracy while also providing new evidence about the relationship between climate risk and oil prices.展开更多
Symbolic regression(SR),exploring mathematical expressions from a given data set to construct an interpretable model,emerges as a powerful computational technique with the potential to transform the“black box”machin...Symbolic regression(SR),exploring mathematical expressions from a given data set to construct an interpretable model,emerges as a powerful computational technique with the potential to transform the“black box”machining learning methods into physical and chemistry interpretable expressions in material science research.In this review,the current advancements in SR are investigated,focusing on the underlying theories,fundamental flowcharts,various techniques,implemented codes,and application fields.More predominantly,the challenging issues and future opportunities in SR that should be overcome to unlock the full potential of SR in material design and research,including graphics processing unit accelera-tion and transfer learning algorithms,the trade-off between expression accuracy and complexity,physical or chemistry interpretable SR with generative large language models,and multimodal SR methods,are discussed.展开更多
This work attempts to understand how a customized bus(CB)operator decides to open or close a CB line.We look into the changes in the operation status of CB lines(i.e.reopening and closure)from one of the largest CB op...This work attempts to understand how a customized bus(CB)operator decides to open or close a CB line.We look into the changes in the operation status of CB lines(i.e.reopening and closure)from one of the largest CB operators in Shanghai,China,with a 22-month con secutive observation ranging from January 2019 to October 2020.As all CB services were totally suspended at the beginning of 2020 due to the COVID-19 travel restriction and then gradually recovered in March 2020,we utilize this study period as a naturalistic observa tion experiment to investigate the changes in the operation status of each CB line before and after the travel restriction.Using the operation status at each month as the binary alternatives,the mixed logit models and the tree-based models with explainable machine learning techniques are respectively adopted to explore the factors that influence the decision-making process.The findings from both types of models are in general consistent.The results show that the characteristics of each CB line including the ridership,the length of the line,the closeness to charging stations,and the overlap of CB lines significantly impact the decisions.In addition,the land-use types around the CB stops and the market competition from alternative travel modes also play a key role in making the decisions.展开更多
基金support from China Scholarship Council(CSC)(202406440073).
文摘Accurate prediction of pure component physiochemical properties is crucial for process integration, multiscale modelling, and optimization. In this work, an enhanced framework for pure component property prediction by using explainable machine learning methods is proposed. In this framework, the molecular representation method based on the connectivity matrix effectively considers atomic bonding relationships to automatically generate features. The supervised machine learning model random forest is applied for feature ranking and pooling. The adjusted R^(2) is introduced to penalize the inclusion of additional features, providing an assessment of the true contribution of features. The prediction results for normal boiling point (T_(b)), liquid molar volume (L_(mv)), critical temperature (T_(c)) and critical pressure (P_(c)) obtained using Artificial Neural Network and Gaussian Process Regression models confirm the accuracy of the molecular representation method. Comparison with GC based models shows that the root-mean-square error on the test set can be reduced by up to 83.8%. To enhance the interpretability of the model, a feature analysis method based on Shapley values is employed to determine the contribution of each feature to the property predictions. The results indicate that using the feature pooling method reduces the number of features from 13316 to 100 without compromising model accuracy. The feature analysis results for Tb, Lmv, Tc, and Pc confirms that different molecular properties are influenced by different structural features, aligning with mechanistic interpretations. In conclusion, the proposed framework is demonstrated to be feasible and provides a solid foundation for mixture component reconstruction and process integration modelling.
基金funding from the National Key Research and Development Program of China(2023YFC3206300)the Gansu Provincial Science and Technology Program(22ZD6FA005)+2 种基金the Gansu Youth Science and Technology Fund(E4310103)the Gansu Postdoctoral Science Foundation(E339880112)the Tibet Science and Technology Program(XZ202301ZY0001G and XZ202401JD0007)。
文摘Glacier mass balance is a key indicator of glacier health and climate change sensitivity.Influencing factors include both climatic and nonclimatic elements,forming a complex set of drivers.There is a lack of quantitative analysis of these composite factors,particularly in climate-typical regions like the Tanggula Mountains on the central Tibetan Plateau.We collected data on various factors affecting glacier mass balance from 2000 to 2020,including climate variables,topographic variables,geometric parameters,and glacier dynamics.We utilized linear regression models,ensemble learning models,and Open Global Glacier Model(OGGM)to analyze glacier mass balance changes in the Tanggula Mountains.Results indicate that linear models explain 58%of the variance in glacier mass balance,with seasonal temperature and precipitation having significant impacts.Our findings show that ensemble learning models made the explanations 5.2%more accurate by including the impact of topographic and geometric factors such as the average glacier height,the slope of the glacier tongue,the speed of the ice flow,and the area of the glacier.Interpretable machine learning identified the spatial distribution of positive and negative impacts of these characteristics and the interaction between glacier topography and ice dynamics.Finally,we predicted the responses of glaciers of different sizes to future climate change based on the results of interpretable machine learning.It was found that relatively large glaciers(>1 km~2)are likely to persist until the end of this century under low emission scenarios,whereas small glaciers(<1 km~2)are expected to nearly disappear by 2080 under any emission scenario.Our research provides technical support for improving glacier change modeling and protection on the Tibetan Plateau.
文摘Background:Frailty in older adults is linked to increased risks and lower quality of life.Pre-frailty,a condition preceding frailty,is intervenable,but its determinants and assessment are challenging.This study aims to develop and validate an explainable machine learning model for pre-frailty risk assessment among community-dwelling older adults.Methods:The study included 3141 adults aged 60 or above from the China Health and Retirement Longitudinal Study.Pre-frailty was characterized by one or two criteria from the physical frailty phenotype scale.We extracted 80 distinct features across seven dimensions to evaluate pre-frailty risk.A model was constructed using recursive feature elimination and a stacking-CatBoost distillation module on 80%of the sample and validated on a separate 20%holdout data set.Results:The study used data from 2508 community-dwelling older adults(mean age,67.24 years[range,60–96];1215[48.44%]females)to develop a pre-frailty risk assessment model.We selected 57 predictive features and built a distilled CatBoost model,which achieved the highest discrimination(AUROC:0.7560[95%CI:0.7169,0.7928])on the 20%holdout data set.The living city,BMI,and peak expiratory flow(PEF)were the three most significant contributors to pre-frailty risk.Physical and environmental factors were the top 2 impactful feature dimensions.Conclusions:An accurate and interpretable pre-frailty risk assessment framework using state-of-the-art machine learning techniques and explanation methods has been developed.Our framework incorporates a wide range of features and determinants,allowing for a comprehensive and nuanced understanding of pre-frailty risk.
基金supported by Princess Nourah bint Abdulrahman University Researchers Supporting Project number(PNURSP2022R196),Princess Nourah bint Abdulrahman University,Riyadh,Saudi Arabia.
文摘Nowadays,quantum machine learning is attracting great interest in a wide range offields due to its potential superior performance and capabilities.The massive increase in computational capacity and speed of quantum computers can lead to a quantum leap in the healthcarefield.Heart disease seriously threa-tens human health since it is the leading cause of death worldwide.Quantum machine learning methods can propose effective solutions to predict heart disease and aid in early diagnosis.In this study,an ensemble machine learning model based on quantum machine learning classifiers is proposed to predict the risk of heart disease.The proposed model is a bagging ensemble learning model where a quantum support vector classifier was used as a base classifier.Further-more,in order to make the model’s outcomes more explainable,the importance of every single feature in the prediction is computed and visualized using SHapley Additive exPlanations(SHAP)framework.In the experimental study,other stand-alone quantum classifiers,namely,Quantum Support Vector Classifier(QSVC),Quantum Neural Network(QNN),and Variational Quantum Classifier(VQC)are applied and compared with classical machine learning classifiers such as Sup-port Vector Machine(SVM),and Artificial Neural Network(ANN).The experi-mental results on the Cleveland dataset reveal the superiority of QSVC compared to the others,which explains its use in the proposed bagging model.The Bagging-QSVC model outperforms all aforementioned classifiers with an accuracy of 90.16%while showing great competitiveness compared to some state-of-the-art models using the same dataset.The results of the study indicate that quantum machine learning classifiers perform better than classical machine learning classi-fiers in predicting heart disease.In addition,the study reveals that the bagging ensemble learning technique is effective in improving the prediction accuracy of quantum classifiers.
文摘Retaining walls are utilized to support the earth and prevent the soil from spreading with natural slope angles where there are differences in the elevation of ground surfaces.As the need for retaining structures increases,the use of retaining walls is increasing.The retaining walls,which increase the stability of levels,are economical and meet existing adverse conditions.A considerable amount of retaining walls is made from steel-reinforced concrete.The construction of reinforced concrete retaining walls can be costly due to its components.For this reason,the optimum cost should be targeted in the design of retaining walls.This study presents an artificial neural network(ANN)model developed to predict the optimum dimensions of a retaining wall using soil properties,material properties,and external loading conditions.The dataset utilized to train the ANN model is generated with the Flower Pollination Algorithm.The target variables in the dataset are the length of the heel(y1),length of the toe(y2),thickness of the stem(top)(y3),thickness of the stem(bottom)(y4),foundation base thickness(y5)and cost(y6)and these are estimated by utilizing an ANN model based on the height of the wall(x1),material unit weight(x2),wall friction angle(x3),surcharge load(x4),concrete cost per m3(x5),steel cost per ton(x6)and the soil class(x7).The model is formulated and trained as a multi-output regression model,as all outputs are numeric and continuous.The training and evaluation of the model results in a high prediction performance(R20.99).In addition,the impacts of different input features on the model>predictions are revealed using the SHapley Additive exPlanations(SHAP)algorithm.The study demonstrates that when trained with a large dataset,ANN models perform very well by predicting the optimal cost with high performance.
文摘As more medical data become digitalized,machine learning is regarded as a promising tool for constructing medical decision support systems.Even with vast medical data volumes,machine learning is still not fully exploiting its potential because the data usually sits in data silos,and privacy and security regulations restrict their access and use.To address these issues,we built a secured and explainable machine learning framework,called explainable federated XGBoost(EXPERTS),which can share valuable information among different medical institutions to improve the learning results without sharing the patients’ data.It also reveals how the machine makes a decision through eigenvalues to offer a more insightful answer to medical professionals.To study the performance,we evaluate our approach by real-world datasets,and our approach outperforms the benchmark algorithms under both federated learning and non-federated learning frameworks.
基金funded by the Postgraduate Research&Practice Innovation Program of Jiangsu Province(SJCX23-2117).
文摘This study develops an approach consisting of a stacking model integrated with a multi-objective optimisation algorithm aimed at predicting and optimising the ecological performance of buildings.The integrated model consists of five base models and a meta-model,which significantly improves the prediction performance.Specifically,the R2 value was improved by 9.19% and the error metrics MAE,MSE,MAPE,and CVRMSE were reduced by 69.47%,79.88%,67.32%,and 57.02%,respectively,compared to the single prediction model.According to the research on interpretable machine learning,adding the SHAP value gives us a deeper understanding of the impact of each architectural design parameter on the performance.In the multi-objective optimisation part,we used the NSGA-Ⅲ algorithm to successfully improve the energy efficiency,daylight utilisation and thermal comfort of the building.Specifically,the optimal design solution reduces the energy use intensity by 31.6 kWh/m^(2),improves the useful daylight index by 39%,and modulated the thermal comfort index,resulting in a decrement of 0.69℃ for the summer season and an enhancement of 0.64℃ for the winter season,respectively.Overall,this study provides building designers and decision makers with a tool to make better design decisions at an early stage to achieve a better combination of energy efficiency,daylight utilisation and thermal comfort optimisation in an integrated manner,providing an important support for achieving sustainable building design.
基金supported by the National Natural Science Foundation of China(Grant No.72074111).
文摘Participants in oil markets are increasingly aware of the climate risks posed by frequent extreme weather.This paper examines the role of extremely high-temperature weather information in predicting oil futures prices on the China International Energy Exchange(INE).An extreme high-temperature weather index(HTI)is developed on the basis of meteorological data at INE’s crude oil production and storage sites.The local interpretable model-agnostic explanations(LIME)and accumulated local effects(ALE)methods are used to compare the predictive contribution of the HTI with that of 15 common predictors.The results indicate that the HTI enhances the out-of-sample accuracy of five classical prediction models for INE oil prices.The recurrent neural network(RNN)model exhibits superior out-of-sample forecast performance,with an MAE of 14.379,an RMSE of 19.624,and a DS of 66.67%.The predictive importance of the HTI in the best RNN model ranks third in most test instances,surpassing conventional oil price predictors such as stock market indicators.The ALE analysis reveals a positive correlation between extremely high-temperature weather and INE oil prices.These findings can help investors and oil market regulators improve oil price forecast accuracy while also providing new evidence about the relationship between climate risk and oil prices.
基金National Natural Science Foundation of China,Grant/Award Number:52332005National Key Research and Development Program of China,Grant/Award Number:2022YFB3807200China Postdoctoral Science Foundation,Grant/Award Number:2022TQ0019。
文摘Symbolic regression(SR),exploring mathematical expressions from a given data set to construct an interpretable model,emerges as a powerful computational technique with the potential to transform the“black box”machining learning methods into physical and chemistry interpretable expressions in material science research.In this review,the current advancements in SR are investigated,focusing on the underlying theories,fundamental flowcharts,various techniques,implemented codes,and application fields.More predominantly,the challenging issues and future opportunities in SR that should be overcome to unlock the full potential of SR in material design and research,including graphics processing unit accelera-tion and transfer learning algorithms,the trade-off between expression accuracy and complexity,physical or chemistry interpretable SR with generative large language models,and multimodal SR methods,are discussed.
基金This work is supported by the National Natural Science Foundation of China(Grant Nos.52272321,71901164,52272335)the Fundamental Research Funds for the Central Universities.
文摘This work attempts to understand how a customized bus(CB)operator decides to open or close a CB line.We look into the changes in the operation status of CB lines(i.e.reopening and closure)from one of the largest CB operators in Shanghai,China,with a 22-month con secutive observation ranging from January 2019 to October 2020.As all CB services were totally suspended at the beginning of 2020 due to the COVID-19 travel restriction and then gradually recovered in March 2020,we utilize this study period as a naturalistic observa tion experiment to investigate the changes in the operation status of each CB line before and after the travel restriction.Using the operation status at each month as the binary alternatives,the mixed logit models and the tree-based models with explainable machine learning techniques are respectively adopted to explore the factors that influence the decision-making process.The findings from both types of models are in general consistent.The results show that the characteristics of each CB line including the ridership,the length of the line,the closeness to charging stations,and the overlap of CB lines significantly impact the decisions.In addition,the land-use types around the CB stops and the market competition from alternative travel modes also play a key role in making the decisions.
