Neural network ensemble based on rough sets reduct is proposed to decrease the computational complexity of conventional ensemble feature selection algorithm. First, a dynamic reduction technology combining genetic alg...Neural network ensemble based on rough sets reduct is proposed to decrease the computational complexity of conventional ensemble feature selection algorithm. First, a dynamic reduction technology combining genetic algorithm with resampling method is adopted to obtain reducts with good generalization ability. Second, Multiple BP neural networks based on different reducts are built as base classifiers. According to the idea of selective ensemble, the neural network ensemble with best generalization ability can be found by search strategies. Finally, classification based on neural network ensemble is implemented by combining the predictions of component networks with voting. The method has been verified in the experiment of remote sensing image and five UCI datasets classification. Compared with conventional ensemble feature selection algorithms, it costs less time and lower computing complexity, and the classification accuracy is satisfactory.展开更多
Work exchange is a promising innovative technology in recovering residual pressure energy. However, at the systematic level, the comprehensive utilization of different energy resources in an energy system has become a...Work exchange is a promising innovative technology in recovering residual pressure energy. However, at the systematic level, the comprehensive utilization of different energy resources in an energy system has become an issue of concern. In this work, a systematic approach is proposed, one that successively integrates heat, work and adjusts operation parameters. A detailed procedure for building a heat-work coupling transfer network is provided. The synthesis mainly consists of constructing a work exchange sub-network with pinch analysis based on positive displacement type work exchangers. Simultaneously, another kind of sub-network based on turbine-type work exchangers is built as a schematic comparison. The influence of applying a positive displacement work exchanger on the system is investigated. Finally, as a case study, a renovation design of a typical rectisol process in the coal-water slurry gasification section of an ammonia plant is presented. The results show that the added work exchanger has little impact on the existing heat exchange sub-network. Moreover,extra pressure energy is recovered by coupling the transfer network. It is concluded that the heat-work systematic design is a promising and powerful method to use energy more efficiently.展开更多
Defect engineering is in the limelight for the fabrication of electrochemical energy storage devices.However,determining the influence of the defect density and location on the electrochemical behavior remains challen...Defect engineering is in the limelight for the fabrication of electrochemical energy storage devices.However,determining the influence of the defect density and location on the electrochemical behavior remains challenging.Herein,self-organized TiO_(2)nanotube arrays(TNTAs)are synthesized by anodization,and their oxygen defect location and density are tuned by a controllable post-annealing process.TNTAs annealed at 600℃ in N2 exhibit the highest capacity(289.2 m Ah g^(-1)at 0.8 C)for lithium-ion storage,while those annealed at 900℃ in N2 show a specific capacitance of 35.6 m F cm^(-2)and stability above96%after 10,000 cycles for supercapacitor.Ex situ electron paramagnetic resonance spectra show that the surface-exposed oxygen defects increase,but the bulk embedded oxygen defects decrease with increasing annealing temperature.Density functional theory simulations reveal that a higher density of bulk oxygen defects corresponds to higher localized electrons states,which upshift the Fermi level and facilitate the lithium intercalation kinetic process.Meanwhile,differential charge density calculation indicates that the increase of surface oxygen defects in the anatase(101)plane leads to higher density excess electrons,which act as negative charge centers to enhance the surface potential for ion adsorption.This oxygen-deficient location and density tunable strategy introduce new opportunities for high-energy and high-power-density energy storage systems.展开更多
Background: Neuroblastoma is one common pediatric malignancy notorious forhigh temporal and spatial heterogeneities. More than half of its patients developdistant metastases involving vascularized organs, especially t...Background: Neuroblastoma is one common pediatric malignancy notorious forhigh temporal and spatial heterogeneities. More than half of its patients developdistant metastases involving vascularized organs, especially the bone marrow. It isthus necessary to have an economical, noninvasive method without muchradiation for follow‐ups. Radiomics has been used in many cancers to assistaccurate diagnosis but not yet in bone marrow metastasis in neuroblastoma.Methods: A total of 182 patients with neuroblastoma were retrospectivelycollected and randomly divided into the training and validation sets. Fivehundredand seventy‐two radiomics features were extracted from magneticresonance imaging, among which 41 significant ones were selected via T‐testfor model development. We attempted 13 machine‐learning algorithms andeventually chose three best‐performed models. The integrative performanceevaluations are based on the area under the curves (AUCs), calibration curves,risk deciles plots, and other indexes.Results: Extreme gradient boosting, random forest (RF), and adaptiveboosting were the top three to predict bone marrow metastases in neuroblastoma while RF was the most accurate one. Its AUC was 0.90(0.86–0.93), F1 score was 0.82, sensitivity was 0.76, and negative predictivevalue was 0.79 in the training set. The values were 0.82 (0.71–0.93), 0.80,0.75, and 0.92 in the validation set, respectively.Conclusions: Radiomics models are likely to contribute more to metastaticdiagnoses and the formulation of personalized healthcare strategies in clinics.It has great potential of being a revolutionary method to replace traditionalinterventions in the future.展开更多
基金supported by the National High-Tech Research and Development Plan of China (No.2007AA04Z224)the National Natural Science Foundation of China (No.60775047, 60835004)
文摘Neural network ensemble based on rough sets reduct is proposed to decrease the computational complexity of conventional ensemble feature selection algorithm. First, a dynamic reduction technology combining genetic algorithm with resampling method is adopted to obtain reducts with good generalization ability. Second, Multiple BP neural networks based on different reducts are built as base classifiers. According to the idea of selective ensemble, the neural network ensemble with best generalization ability can be found by search strategies. Finally, classification based on neural network ensemble is implemented by combining the predictions of component networks with voting. The method has been verified in the experiment of remote sensing image and five UCI datasets classification. Compared with conventional ensemble feature selection algorithms, it costs less time and lower computing complexity, and the classification accuracy is satisfactory.
基金supported by the National Natural Science Foundation of China (No. 20936004 and No. 21376187)
文摘Work exchange is a promising innovative technology in recovering residual pressure energy. However, at the systematic level, the comprehensive utilization of different energy resources in an energy system has become an issue of concern. In this work, a systematic approach is proposed, one that successively integrates heat, work and adjusts operation parameters. A detailed procedure for building a heat-work coupling transfer network is provided. The synthesis mainly consists of constructing a work exchange sub-network with pinch analysis based on positive displacement type work exchangers. Simultaneously, another kind of sub-network based on turbine-type work exchangers is built as a schematic comparison. The influence of applying a positive displacement work exchanger on the system is investigated. Finally, as a case study, a renovation design of a typical rectisol process in the coal-water slurry gasification section of an ammonia plant is presented. The results show that the added work exchanger has little impact on the existing heat exchange sub-network. Moreover,extra pressure energy is recovered by coupling the transfer network. It is concluded that the heat-work systematic design is a promising and powerful method to use energy more efficiently.
基金supported by the National Nature Science Foundation of China(11575025,U1832176)the Science and Technology Project of Beijing(Z171100002017008)the Fundamental Research Funds for the Central Universities。
文摘Defect engineering is in the limelight for the fabrication of electrochemical energy storage devices.However,determining the influence of the defect density and location on the electrochemical behavior remains challenging.Herein,self-organized TiO_(2)nanotube arrays(TNTAs)are synthesized by anodization,and their oxygen defect location and density are tuned by a controllable post-annealing process.TNTAs annealed at 600℃ in N2 exhibit the highest capacity(289.2 m Ah g^(-1)at 0.8 C)for lithium-ion storage,while those annealed at 900℃ in N2 show a specific capacitance of 35.6 m F cm^(-2)and stability above96%after 10,000 cycles for supercapacitor.Ex situ electron paramagnetic resonance spectra show that the surface-exposed oxygen defects increase,but the bulk embedded oxygen defects decrease with increasing annealing temperature.Density functional theory simulations reveal that a higher density of bulk oxygen defects corresponds to higher localized electrons states,which upshift the Fermi level and facilitate the lithium intercalation kinetic process.Meanwhile,differential charge density calculation indicates that the increase of surface oxygen defects in the anatase(101)plane leads to higher density excess electrons,which act as negative charge centers to enhance the surface potential for ion adsorption.This oxygen-deficient location and density tunable strategy introduce new opportunities for high-energy and high-power-density energy storage systems.
基金Chinese Postdoctoral Natural Funding,Grant/Award Number:2022M710884Research Foundation of Guangzhou Women and Children's Medical Center,Grant/Award Number:KTa377a204193688National Natural Science Foundation of China,Grant/Award Number:82202251。
文摘Background: Neuroblastoma is one common pediatric malignancy notorious forhigh temporal and spatial heterogeneities. More than half of its patients developdistant metastases involving vascularized organs, especially the bone marrow. It isthus necessary to have an economical, noninvasive method without muchradiation for follow‐ups. Radiomics has been used in many cancers to assistaccurate diagnosis but not yet in bone marrow metastasis in neuroblastoma.Methods: A total of 182 patients with neuroblastoma were retrospectivelycollected and randomly divided into the training and validation sets. Fivehundredand seventy‐two radiomics features were extracted from magneticresonance imaging, among which 41 significant ones were selected via T‐testfor model development. We attempted 13 machine‐learning algorithms andeventually chose three best‐performed models. The integrative performanceevaluations are based on the area under the curves (AUCs), calibration curves,risk deciles plots, and other indexes.Results: Extreme gradient boosting, random forest (RF), and adaptiveboosting were the top three to predict bone marrow metastases in neuroblastoma while RF was the most accurate one. Its AUC was 0.90(0.86–0.93), F1 score was 0.82, sensitivity was 0.76, and negative predictivevalue was 0.79 in the training set. The values were 0.82 (0.71–0.93), 0.80,0.75, and 0.92 in the validation set, respectively.Conclusions: Radiomics models are likely to contribute more to metastaticdiagnoses and the formulation of personalized healthcare strategies in clinics.It has great potential of being a revolutionary method to replace traditionalinterventions in the future.
