Decentralized machine learning frameworks,e.g.,federated learning,are emerging to facilitate learning with medical data under privacy protection.It is widely agreed that the establishment of an accurate and robust med...Decentralized machine learning frameworks,e.g.,federated learning,are emerging to facilitate learning with medical data under privacy protection.It is widely agreed that the establishment of an accurate and robust medical learning model requires a large number of continuous synchronous monitoring data of patients from various types of monitoring facilities.However,the clinic monitoring data are usually sparse and imbalanced with errors and time irregularity,leading to inaccurate risk prediction results.To address this issue,this paper designs a medical data resampling and balancing scheme for federated learning to eliminate model biases caused by sample imbalance and provide accurate disease risk prediction on multi-center medical data.Experimental results on a real-world clinical database MIMIC-Ⅳ demonstrate that the proposed method can improve AUC(the area under the receiver operating characteristic) from 50.1% to 62.8%,with a significant performance improvement of accuracy from 76.8% to 82.2%,compared to a vanilla federated learning artificial neural network(ANN).Moreover,we increase the model’s tolerance for missing data from 20% to 50% compared with a stand-alone baseline model.展开更多
Real-time prediction of the rock mass class in front of the tunnel face is essential for the adaptive adjustment of tunnel boring machines(TBMs).During the TBM tunnelling process,a large number of operation data are g...Real-time prediction of the rock mass class in front of the tunnel face is essential for the adaptive adjustment of tunnel boring machines(TBMs).During the TBM tunnelling process,a large number of operation data are generated,reflecting the interaction between the TBM system and surrounding rock,and these data can be used to evaluate the rock mass quality.This study proposed a stacking ensemble classifier for the real-time prediction of the rock mass classification using TBM operation data.Based on the Songhua River water conveyance project,a total of 7538 TBM tunnelling cycles and the corresponding rock mass classes are obtained after data preprocessing.Then,through the tree-based feature selection method,10 key TBM operation parameters are selected,and the mean values of the 10 selected features in the stable phase after removing outliers are calculated as the inputs of classifiers.The preprocessed data are randomly divided into the training set(90%)and test set(10%)using simple random sampling.Besides stacking ensemble classifier,seven individual classifiers are established as the comparison.These classifiers include support vector machine(SVM),k-nearest neighbors(KNN),random forest(RF),gradient boosting decision tree(GBDT),decision tree(DT),logistic regression(LR)and multilayer perceptron(MLP),where the hyper-parameters of each classifier are optimised using the grid search method.The prediction results show that the stacking ensemble classifier has a better performance than individual classifiers,and it shows a more powerful learning and generalisation ability for small and imbalanced samples.Additionally,a relative balance training set is obtained by the synthetic minority oversampling technique(SMOTE),and the influence of sample imbalance on the prediction performance is discussed.展开更多
With tunnel boring machine being used in underground engineering,accurate geological indicators have been the important basis for tunnel boring machine(TBM)construction.Back propagation neural network(BPNN)has been us...With tunnel boring machine being used in underground engineering,accurate geological indicators have been the important basis for tunnel boring machine(TBM)construction.Back propagation neural network(BPNN)has been used to predict the geological indicators of tunnels in previous studies.Nevertheless,these studies ignored the imbalance proportion of surrounding rock grades,leading to the indiscriminate use of data,thus affecting the predictive effect of BPNN.In order to prove the importance of the proportion of surround-ing rock grade in geological prediction,we mainly attempt to utilize particle swarm optimization(PSO)to optimize the proportion of sample data,and integrate with BPNN to establish a PSO-BPNN theoretical model to predict geological indicators.At the same time,combined with the actual engineering data,5 tunneling indicators were selected as input and 4 geological indicators were selected as out-put by a variety of dimensionality reduction methods.The geological indicators are density,uniaxial compressive strength,internal fric-tion angle(u)and Poisson’s ratio(e).On this basis,the PSO-BPNN prediction model was established in detail.By comparing the prediction of traditional BPNN,PSO-BPNN and other optimization-integrated models,the result shows that optimized proportion of surrounding rock grades reduces the prediction error and improves the interpretability of the prediction model.Meanwhile,we com-bined the theory of surrounding rock partition to illustrate the rationality of surrounding rock proportion in PSO result,that is,the proportion of complex surrounding rock should be increased appropriately to improve the prediction result.Ultimately,based on the optimization-integrated models with engineering data and the surrounding rock classification theory,the importance of proportion of surrounding rock grades for tunnel geological prediction is confirmed.展开更多
基金supported by Hubei Provincial Development and Reform Commission Program"Hubei Big Data Analysis Platform and Intelligent Service Project for Medical and Health"。
文摘Decentralized machine learning frameworks,e.g.,federated learning,are emerging to facilitate learning with medical data under privacy protection.It is widely agreed that the establishment of an accurate and robust medical learning model requires a large number of continuous synchronous monitoring data of patients from various types of monitoring facilities.However,the clinic monitoring data are usually sparse and imbalanced with errors and time irregularity,leading to inaccurate risk prediction results.To address this issue,this paper designs a medical data resampling and balancing scheme for federated learning to eliminate model biases caused by sample imbalance and provide accurate disease risk prediction on multi-center medical data.Experimental results on a real-world clinical database MIMIC-Ⅳ demonstrate that the proposed method can improve AUC(the area under the receiver operating characteristic) from 50.1% to 62.8%,with a significant performance improvement of accuracy from 76.8% to 82.2%,compared to a vanilla federated learning artificial neural network(ANN).Moreover,we increase the model’s tolerance for missing data from 20% to 50% compared with a stand-alone baseline model.
基金funded by the National Natural Science Foundation of China(Grant No.41941019)the State Key Laboratory of Hydroscience and Engineering(Grant No.2019-KY-03)。
文摘Real-time prediction of the rock mass class in front of the tunnel face is essential for the adaptive adjustment of tunnel boring machines(TBMs).During the TBM tunnelling process,a large number of operation data are generated,reflecting the interaction between the TBM system and surrounding rock,and these data can be used to evaluate the rock mass quality.This study proposed a stacking ensemble classifier for the real-time prediction of the rock mass classification using TBM operation data.Based on the Songhua River water conveyance project,a total of 7538 TBM tunnelling cycles and the corresponding rock mass classes are obtained after data preprocessing.Then,through the tree-based feature selection method,10 key TBM operation parameters are selected,and the mean values of the 10 selected features in the stable phase after removing outliers are calculated as the inputs of classifiers.The preprocessed data are randomly divided into the training set(90%)and test set(10%)using simple random sampling.Besides stacking ensemble classifier,seven individual classifiers are established as the comparison.These classifiers include support vector machine(SVM),k-nearest neighbors(KNN),random forest(RF),gradient boosting decision tree(GBDT),decision tree(DT),logistic regression(LR)and multilayer perceptron(MLP),where the hyper-parameters of each classifier are optimised using the grid search method.The prediction results show that the stacking ensemble classifier has a better performance than individual classifiers,and it shows a more powerful learning and generalisation ability for small and imbalanced samples.Additionally,a relative balance training set is obtained by the synthetic minority oversampling technique(SMOTE),and the influence of sample imbalance on the prediction performance is discussed.
基金supported by the National Natural Science Foundation of China,China(Grant No.52075370).
文摘With tunnel boring machine being used in underground engineering,accurate geological indicators have been the important basis for tunnel boring machine(TBM)construction.Back propagation neural network(BPNN)has been used to predict the geological indicators of tunnels in previous studies.Nevertheless,these studies ignored the imbalance proportion of surrounding rock grades,leading to the indiscriminate use of data,thus affecting the predictive effect of BPNN.In order to prove the importance of the proportion of surround-ing rock grade in geological prediction,we mainly attempt to utilize particle swarm optimization(PSO)to optimize the proportion of sample data,and integrate with BPNN to establish a PSO-BPNN theoretical model to predict geological indicators.At the same time,combined with the actual engineering data,5 tunneling indicators were selected as input and 4 geological indicators were selected as out-put by a variety of dimensionality reduction methods.The geological indicators are density,uniaxial compressive strength,internal fric-tion angle(u)and Poisson’s ratio(e).On this basis,the PSO-BPNN prediction model was established in detail.By comparing the prediction of traditional BPNN,PSO-BPNN and other optimization-integrated models,the result shows that optimized proportion of surrounding rock grades reduces the prediction error and improves the interpretability of the prediction model.Meanwhile,we com-bined the theory of surrounding rock partition to illustrate the rationality of surrounding rock proportion in PSO result,that is,the proportion of complex surrounding rock should be increased appropriately to improve the prediction result.Ultimately,based on the optimization-integrated models with engineering data and the surrounding rock classification theory,the importance of proportion of surrounding rock grades for tunnel geological prediction is confirmed.