In this paper,a novel train positioning method considering satellite raw observation data was proposed,which aims to promote train positioning performance from an innovative perspective of the train satellite-based po...In this paper,a novel train positioning method considering satellite raw observation data was proposed,which aims to promote train positioning performance from an innovative perspective of the train satellite-based positioning error sources.The method focused on overcoming the abnormal observations in satellite observation data caused by railway environment rather than the positioning results.Specifically,the relative positioning experimental platform was built and the zero-baseline method was firstly employed to evaluate the carrier phase data quality,and then,GNSS combined observation models were adopted to construct the detection values,which were applied to judge abnormal-data through the dual-frequency observations.Further,ambiguity fixing optimization was investigated based on observation data selection in partly-blocked environments.The results show that the proposed method can effectively detect and address abnormal observations and improve positioning stability.Cycle slips and gross errors can be detected and identified based on dual-frequency global navigation satellite system data.After adopting the data selection strategy,the ambiguity fixing percentage was improved by 29.2%,and the standard deviation in the East,North,and Up components was enhanced by 12.7%,7.4%,and 12.5%,respectively.The proposed method can provide references for train positioning performance optimization in railway environments from the perspective of positioning error sources.展开更多
软测量技术为工业过程中重要变量及难测变量的预测提供了一个有效的解决办法。然而,由于工业过程的复杂化和高昂的数据获取成本,使得标记数据与未标记数据分布不平衡。此时,构建高性能的软测量模型成为一个挑战。针对这一问题,提出了一...软测量技术为工业过程中重要变量及难测变量的预测提供了一个有效的解决办法。然而,由于工业过程的复杂化和高昂的数据获取成本,使得标记数据与未标记数据分布不平衡。此时,构建高性能的软测量模型成为一个挑战。针对这一问题,提出了一种基于时差的多输出tri-training异构软测量方法。通过构建一种新的tri-training框架,采用多输出的高斯过程回归(multi-output Gaussian process regression,MGPR)、相关向量机(multi-output relevance vector machine,MRVM)、最小二乘支持向量机(multi-output least squares support vector machine,MLSSVM)三种模型作为基线监督回归器,使用标记数据进行训练和迭代;同时,引入时间差分(time difference,TD)改进模型的动态特性,并通过卡尔曼滤波(Kalman filtering,KF)优化模型的参数,提高其预测性能;最后通过模拟污水处理平台(benchmark simulation model 1,BSM1)和实际污水处理厂对该模型进行了验证。结果表明,与传统的软测量建模方法相比,该模型能显著提高数据分布不平衡下软测量模型的自适应性和预测性能。展开更多
Ventilation systems are critical for improving the cabin environment in high-speed trains,and their interest has increased significantly.However,whether air supply non-verticality deteriorates the cabin air environmen...Ventilation systems are critical for improving the cabin environment in high-speed trains,and their interest has increased significantly.However,whether air supply non-verticality deteriorates the cabin air environment,and the flow mechanism behind it and the degree of deterioration are not known.This study first analyzes the interaction between deflection angle and cabin flow field characteristics and ventilation performance.The results revealed that the interior temperature and pollutant concentration decreased slightly with increasing deflection angle,but resulted in significant deterioration of thermal comfort and air quality.This is evidenced by an increase in both draught rate and non-uniformity coefficient,an increase in the number of measurement points that do not satisfy the micro-wind speed and temperature difference requirements by about 5% and 15%,respectively,and an increase in longitudinal penetration of pollutants by a factor of about 5 and the appearance of locking regions at the ends of cabin.The results also show that changing the deflection pattern only affects the region of deterioration and does not essentially improve this deterioration.This study can provide reference and help for the ventilation design of high-speed trains.展开更多
Irregularities in the track and uneven forces acting on the train can cause shifts in the position of the superconducting magnetic levitation train relative to the track during operation.These shifts lead to asymmetri...Irregularities in the track and uneven forces acting on the train can cause shifts in the position of the superconducting magnetic levitation train relative to the track during operation.These shifts lead to asymmetries in the flow field structure on both sides of the narrow suspension gap,resulting in instability and deterioration of the train’s aerodynamic characteristics,significantly impacting its operational safety.In this study,we firstly validate the aerodynamic characteristics of the superconducting magnetic levitation system by developing a numerical simulation method based on wind tunnel test results.We then investigate the influence of lateral translation parameters on the train’s aerodynamic performance under conditions both with and without crosswinds.We aim to clarify the evolution mechanism of the flow field characteristics under the coupling effect between the train and the U-shaped track and to identify the most unfavorable operational parameters contributing to the deterioration of the train’s aerodynamic properties.The findings show that,without crosswinds,a lateral translation of 30 mm causes a synchronous resonance phenomenon at the side and bottom gaps of the train-track coupling,leading to the worst aerodynamic performance.Under crosswind conditions,a lateral translation of 40 mm maximizes peak pressure fluctuations and average turbulent kinetic energy around the train,resulting in the poorest aerodynamic performance.This research provides theoretical support for enhancing the operational stability of superconducting magnetic levitation trains.展开更多
Bridges crossing active faults are more likely to suffer serious damage or even collapse due to the wreck capabilities of near-fault pulses and surface ruptures under earthquakes.Taking a high-speed railway simply-sup...Bridges crossing active faults are more likely to suffer serious damage or even collapse due to the wreck capabilities of near-fault pulses and surface ruptures under earthquakes.Taking a high-speed railway simply-supported girder bridge with eight spans crossing an active strike-slip fault as the research object,a refined coupling dynamic model of the high-speed train-CRTS III slab ballastless track-bridge system was established based on ABAQUS.The rationality of the established model was thoroughly discussed.The horizontal ground motions in a fault rupture zone were simulated and transient dynamic analyses of the high-speed train-track-bridge coupling system under 3-dimensional seismic excitations were subsequently performed.The safe running speed limits of a high-speed train under different earthquake levels(frequent occurrence,design and rare occurrence)were assessed based on wheel-rail dynamic(lateral wheel-rail force,derailment coefficient and wheel-load reduction rate)and rail deformation(rail dislocation,parallel turning angle and turning angle)indicators.Parameter optimization was then investigated in terms of the rail fastener stiffness and isolation layer friction coefficient.Results of the wheel-rail dynamic indicators demonstrate the safe running speed limits for the high-speed train to be approximately 200 km/h and 80 km/h under frequent and design earthquakes,while the train is unable to run safely under rare earthquakes.In addition,the rail deformations under frequent,design and rare earthquakes meet the safe running requirements of the high-speed train for the speeds of 250,100 and 50 km/h,respectively.The speed limits determined for the wheel-rail dynamic indicators are lower due to the complex coupling effect of the train-track-bridge system under track irregularity.The running safety of the train was improved by increasing the fastener stiffness and isolation layer friction coefficient.At the rail fastener lateral stiffness of 60 kN/mm and isolation layer friction coefficients of 0.9 and 0.8,respectively,the safe running speed limits of the high-speed train increased to 250 km/h and 100 km/h under frequent and design earthquakes,respectively.展开更多
针对复杂交通场景车辆检测算法自适应能力差的问题,提出了基于Co-training半监督学习方法的车辆鲁棒检测算法.首先,针对手工标记的少量样本,分别训练基于Haar-like特征的AdaBoost分类器和基于HOG(histograms of oriented gradients)特征...针对复杂交通场景车辆检测算法自适应能力差的问题,提出了基于Co-training半监督学习方法的车辆鲁棒检测算法.首先,针对手工标记的少量样本,分别训练基于Haar-like特征的AdaBoost分类器和基于HOG(histograms of oriented gradients)特征的SVM(support vector machines)分类器,使其具有一定的识别能力;然后,基于Co-training半监督学习框架,将利用2种算法进行分类得到的新样本分别加入到对方的样本库中,增加训练样本数量,再次进行分类器的训练.由于这2类特征具有冗余性,各自检测出的正负样本包含对方漏检和误检的图像.由于样本数的增加,再次训练所得到的新分类器的鲁棒性得到了很大提高,能更加准确地检测出车辆,而且由算法对未标记样本进行分类标记,不再需要人为标记,提高了车辆检测算法的自适应能力.展开更多
Aerodynamic drag is proportional to the square of speed. With the increase of the speed of train, aerodynamic drag plays an important role for high-speed train. Thus, the reduction of aerodynamic drag and energy consu...Aerodynamic drag is proportional to the square of speed. With the increase of the speed of train, aerodynamic drag plays an important role for high-speed train. Thus, the reduction of aerodynamic drag and energy consumption of high-speed train is one of the essential issues for the development of the desirable train system. Aerodynamic drag on the traveling train is divided into pressure drag and friction one. Pressure drag of train is the force caused by the pressure distribution on the train along the reverse running direction. Friction drag of train is the sum of shear stress, which is the reverse direction of train running direction. In order to reduce the aerodynamic drag, adopting streamline shape of train is the most effective measure. The velocity of the train is related to its length and shape. The outer wind shields can reduce train's air drag by about 15%. At the same time, the train with bottom cover can reduce the air drag by about 50%, compared with the train without bottom plate or skirt structure.展开更多
In this study, experiments were carried out to investigate aerodynamic characteristics of a high-speed train on viaducts in turbulent crosswinds using a 1:25 scaled sectional model wind-tunnel testing. Pressure measur...In this study, experiments were carried out to investigate aerodynamic characteristics of a high-speed train on viaducts in turbulent crosswinds using a 1:25 scaled sectional model wind-tunnel testing. Pressure measurements of two typical sections, one train-head section and one train-body section, at the windward and leeward tracks were conducted under the smooth and turbulence flows with wind attack angles between-6° and 6°, and the corresponding aerodynamic force coefficients were also calculated using the integral method. The experimental results indicate that the track position affects the mean aerodynamic characteristics of the vehicle, especially for the train-body section. The fluctuating pressure coefficients at the leeward track are more significantly affected by the bridge interference compared to those at the windward track. The effect of turbulence on the train-head section is less than that on the train-body section. Additionally, the mean aerodynamic force coefficients are almost negatively correlated to wind attack angles, which is more prominent for vehicles at the leeward track. Moreover, the lateral force plays a critical role in determining the corresponding overturning moment, especially on the train-body section.展开更多
This study investigates the influence of different pantograph parameters and train length on the aerodynamic drag of high-speed train by the delayed detached eddy simulation(DDES) method. The train geometry considered...This study investigates the influence of different pantograph parameters and train length on the aerodynamic drag of high-speed train by the delayed detached eddy simulation(DDES) method. The train geometry considered is the high-speed train with pantographs, and the different versions have 3, 5, 8, 10, 12, 16 and 17 cars. The numerical results are verified by the wind tunnel test with 3.6% difference. The influences of the number of cars and the position, quantity and configuration of pantographs on flow field around high-speed train and wake vortices are analyzed. The aerodynamic drag of middle cars gradually decreases along the flow direction. The aerodynamic drag of pantographs decreases with its backward shift, and that of the first pantograph decreases significantly. As the number of pantographs increases, its effect on the aerodynamic drag decrease of rear cars is more significant. The engineering application equation for the aerodynamic drag of high-speed train with pantographs is proposed. For the 10-car and 17-car train, the differences of total aerodynamic drag between the equation and the simulation results are 1.2% and 0.4%, respectively. The equation generalized in this study could well guide the design phase of high-speed train.展开更多
本文充分利用网页数据的超链接关系和文本信息,提出了一种用于网页分类的归纳式半监督学习算法:基于图的Co-training网页分类算法(Graph based Co-training algorithmfor web page classification),简称GCo-training,并从理论上证明了...本文充分利用网页数据的超链接关系和文本信息,提出了一种用于网页分类的归纳式半监督学习算法:基于图的Co-training网页分类算法(Graph based Co-training algorithmfor web page classification),简称GCo-training,并从理论上证明了算法的有效性.GCo-training在Co-training算法框架下,迭代地学习一个基于由超链接信息构造的图的半监督分类器和一个基于文本特征的Bayes分类器.基于图的半监督分类器只利用少量的标记数据,通过挖掘数据间大量的关系信息就可达到比较高的预测精度,可为Bayes分类器提供大量的标记信息;反过来学习大量标记信息后的Bayes分类器也可为基于图的分类器提供有效信息.迭代过程中,二者互相帮助,不断提高各自的性能,而后Bayes分类器可以用来预测大量未见数据的类别.在Web→KB数据集上的实验结果表明,与利用文本特征和锚文本特征的Co-training算法和基于EM的Bayes算法相比,GCo-training算法性能优越.展开更多
Tri-training能有效利用无标记样例提高泛化能力.针对Tri-training迭代中无标记样例常被错误标记而形成训练集噪声,导致性能不稳定的缺点,文中提出ADE-Tri-training(Tri-training with Adaptive Data Editing)新算法.它不仅利用Remove O...Tri-training能有效利用无标记样例提高泛化能力.针对Tri-training迭代中无标记样例常被错误标记而形成训练集噪声,导致性能不稳定的缺点,文中提出ADE-Tri-training(Tri-training with Adaptive Data Editing)新算法.它不仅利用Remove Only剪辑操作对每次迭代可能产生的误标记样例识别并移除,更重要的是采用自适应策略来确定Remove Only触发与抑制的恰当时机.文中证明,PAC理论下自适应策略中一系列判别充分条件可同时确保新训练集规模迭代增大和新假设分类错误率迭代降低更多.UCI数据集上实验结果表明:ADE-Tri-training具有更好的分类泛化性能和健壮性.展开更多
基金Project(52272339)supported by the National Natural Science Foundation of ChinaProject(2023YFB390730303)supported by the National Key Research and Development Program of China+2 种基金Project(L2023G004)supported by the Science and Technology Research and Development Program of China State Railway Group Co.,Ltd.Project(QZKFKT2023-005)supported by the State Key Laboratory of Heavy-duty and Express High-power Electric Locomotive,ChinaProject(2022JZZ05)supported by the Open Foundation of MOE Key Laboratory of Engineering Structures of Heavy Haul Railway(Central South University),China。
文摘In this paper,a novel train positioning method considering satellite raw observation data was proposed,which aims to promote train positioning performance from an innovative perspective of the train satellite-based positioning error sources.The method focused on overcoming the abnormal observations in satellite observation data caused by railway environment rather than the positioning results.Specifically,the relative positioning experimental platform was built and the zero-baseline method was firstly employed to evaluate the carrier phase data quality,and then,GNSS combined observation models were adopted to construct the detection values,which were applied to judge abnormal-data through the dual-frequency observations.Further,ambiguity fixing optimization was investigated based on observation data selection in partly-blocked environments.The results show that the proposed method can effectively detect and address abnormal observations and improve positioning stability.Cycle slips and gross errors can be detected and identified based on dual-frequency global navigation satellite system data.After adopting the data selection strategy,the ambiguity fixing percentage was improved by 29.2%,and the standard deviation in the East,North,and Up components was enhanced by 12.7%,7.4%,and 12.5%,respectively.The proposed method can provide references for train positioning performance optimization in railway environments from the perspective of positioning error sources.
文摘软测量技术为工业过程中重要变量及难测变量的预测提供了一个有效的解决办法。然而,由于工业过程的复杂化和高昂的数据获取成本,使得标记数据与未标记数据分布不平衡。此时,构建高性能的软测量模型成为一个挑战。针对这一问题,提出了一种基于时差的多输出tri-training异构软测量方法。通过构建一种新的tri-training框架,采用多输出的高斯过程回归(multi-output Gaussian process regression,MGPR)、相关向量机(multi-output relevance vector machine,MRVM)、最小二乘支持向量机(multi-output least squares support vector machine,MLSSVM)三种模型作为基线监督回归器,使用标记数据进行训练和迭代;同时,引入时间差分(time difference,TD)改进模型的动态特性,并通过卡尔曼滤波(Kalman filtering,KF)优化模型的参数,提高其预测性能;最后通过模拟污水处理平台(benchmark simulation model 1,BSM1)和实际污水处理厂对该模型进行了验证。结果表明,与传统的软测量建模方法相比,该模型能显著提高数据分布不平衡下软测量模型的自适应性和预测性能。
基金Project(12372049)supported by the National Natural Science Foundation of ChinaProject(2682023ZTPY036)supported by the Fundamental Research Funds for the Central Universities,ChinaProject(2023TPL-T06)supported by the Independent Project of State Key Laboratory of Rail Transit Vehicle System,China。
文摘Ventilation systems are critical for improving the cabin environment in high-speed trains,and their interest has increased significantly.However,whether air supply non-verticality deteriorates the cabin air environment,and the flow mechanism behind it and the degree of deterioration are not known.This study first analyzes the interaction between deflection angle and cabin flow field characteristics and ventilation performance.The results revealed that the interior temperature and pollutant concentration decreased slightly with increasing deflection angle,but resulted in significant deterioration of thermal comfort and air quality.This is evidenced by an increase in both draught rate and non-uniformity coefficient,an increase in the number of measurement points that do not satisfy the micro-wind speed and temperature difference requirements by about 5% and 15%,respectively,and an increase in longitudinal penetration of pollutants by a factor of about 5 and the appearance of locking regions at the ends of cabin.The results also show that changing the deflection pattern only affects the region of deterioration and does not essentially improve this deterioration.This study can provide reference and help for the ventilation design of high-speed trains.
基金Projects(52372369,52302447,52388102)supported by the National Natural Science Foundation of ChinaProjects(2022YFB4301201-02,2023YFB4302502-02)supported by the National Key R&D Program of China。
文摘Irregularities in the track and uneven forces acting on the train can cause shifts in the position of the superconducting magnetic levitation train relative to the track during operation.These shifts lead to asymmetries in the flow field structure on both sides of the narrow suspension gap,resulting in instability and deterioration of the train’s aerodynamic characteristics,significantly impacting its operational safety.In this study,we firstly validate the aerodynamic characteristics of the superconducting magnetic levitation system by developing a numerical simulation method based on wind tunnel test results.We then investigate the influence of lateral translation parameters on the train’s aerodynamic performance under conditions both with and without crosswinds.We aim to clarify the evolution mechanism of the flow field characteristics under the coupling effect between the train and the U-shaped track and to identify the most unfavorable operational parameters contributing to the deterioration of the train’s aerodynamic properties.The findings show that,without crosswinds,a lateral translation of 30 mm causes a synchronous resonance phenomenon at the side and bottom gaps of the train-track coupling,leading to the worst aerodynamic performance.Under crosswind conditions,a lateral translation of 40 mm maximizes peak pressure fluctuations and average turbulent kinetic energy around the train,resulting in the poorest aerodynamic performance.This research provides theoretical support for enhancing the operational stability of superconducting magnetic levitation trains.
基金Project(51378050) supported by the National Natural Science Foundation of ChinaProject(B13002) supported by the “111” Project,China+2 种基金Project (8192035) supported by the Beijing Municipal Natural Science Foundation,ChinaProject(P2019G002) supported by the Science and Technology Research and Development Program of China RailwayProject(2019YJ193) supported by the State Key Laboratory for Track Technology of High-speed Railway,China。
文摘Bridges crossing active faults are more likely to suffer serious damage or even collapse due to the wreck capabilities of near-fault pulses and surface ruptures under earthquakes.Taking a high-speed railway simply-supported girder bridge with eight spans crossing an active strike-slip fault as the research object,a refined coupling dynamic model of the high-speed train-CRTS III slab ballastless track-bridge system was established based on ABAQUS.The rationality of the established model was thoroughly discussed.The horizontal ground motions in a fault rupture zone were simulated and transient dynamic analyses of the high-speed train-track-bridge coupling system under 3-dimensional seismic excitations were subsequently performed.The safe running speed limits of a high-speed train under different earthquake levels(frequent occurrence,design and rare occurrence)were assessed based on wheel-rail dynamic(lateral wheel-rail force,derailment coefficient and wheel-load reduction rate)and rail deformation(rail dislocation,parallel turning angle and turning angle)indicators.Parameter optimization was then investigated in terms of the rail fastener stiffness and isolation layer friction coefficient.Results of the wheel-rail dynamic indicators demonstrate the safe running speed limits for the high-speed train to be approximately 200 km/h and 80 km/h under frequent and design earthquakes,while the train is unable to run safely under rare earthquakes.In addition,the rail deformations under frequent,design and rare earthquakes meet the safe running requirements of the high-speed train for the speeds of 250,100 and 50 km/h,respectively.The speed limits determined for the wheel-rail dynamic indicators are lower due to the complex coupling effect of the train-track-bridge system under track irregularity.The running safety of the train was improved by increasing the fastener stiffness and isolation layer friction coefficient.At the rail fastener lateral stiffness of 60 kN/mm and isolation layer friction coefficients of 0.9 and 0.8,respectively,the safe running speed limits of the high-speed train increased to 250 km/h and 100 km/h under frequent and design earthquakes,respectively.
基金Project(2001AA505000) supported by the National High-Tech Research and Development of China
文摘Aerodynamic drag is proportional to the square of speed. With the increase of the speed of train, aerodynamic drag plays an important role for high-speed train. Thus, the reduction of aerodynamic drag and energy consumption of high-speed train is one of the essential issues for the development of the desirable train system. Aerodynamic drag on the traveling train is divided into pressure drag and friction one. Pressure drag of train is the force caused by the pressure distribution on the train along the reverse running direction. Friction drag of train is the sum of shear stress, which is the reverse direction of train running direction. In order to reduce the aerodynamic drag, adopting streamline shape of train is the most effective measure. The velocity of the train is related to its length and shape. The outer wind shields can reduce train's air drag by about 15%. At the same time, the train with bottom cover can reduce the air drag by about 50%, compared with the train without bottom plate or skirt structure.
基金Projects(51808563,51925808)supported by the National Natural Science Foundation of ChinaProject(KLWRTBMC18-03)supported by the Open Research Fund of the Key Laboratory of Wind Resistance Technology of Bridges of ChinaProject(2017YFB1201204)supported by the National Key R&D Program of China。
文摘In this study, experiments were carried out to investigate aerodynamic characteristics of a high-speed train on viaducts in turbulent crosswinds using a 1:25 scaled sectional model wind-tunnel testing. Pressure measurements of two typical sections, one train-head section and one train-body section, at the windward and leeward tracks were conducted under the smooth and turbulence flows with wind attack angles between-6° and 6°, and the corresponding aerodynamic force coefficients were also calculated using the integral method. The experimental results indicate that the track position affects the mean aerodynamic characteristics of the vehicle, especially for the train-body section. The fluctuating pressure coefficients at the leeward track are more significantly affected by the bridge interference compared to those at the windward track. The effect of turbulence on the train-head section is less than that on the train-body section. Additionally, the mean aerodynamic force coefficients are almost negatively correlated to wind attack angles, which is more prominent for vehicles at the leeward track. Moreover, the lateral force plays a critical role in determining the corresponding overturning moment, especially on the train-body section.
基金Projects(2018YFB1201801-4,2018YFB1201804-2)supported by National Key R&D Program of China。
文摘This study investigates the influence of different pantograph parameters and train length on the aerodynamic drag of high-speed train by the delayed detached eddy simulation(DDES) method. The train geometry considered is the high-speed train with pantographs, and the different versions have 3, 5, 8, 10, 12, 16 and 17 cars. The numerical results are verified by the wind tunnel test with 3.6% difference. The influences of the number of cars and the position, quantity and configuration of pantographs on flow field around high-speed train and wake vortices are analyzed. The aerodynamic drag of middle cars gradually decreases along the flow direction. The aerodynamic drag of pantographs decreases with its backward shift, and that of the first pantograph decreases significantly. As the number of pantographs increases, its effect on the aerodynamic drag decrease of rear cars is more significant. The engineering application equation for the aerodynamic drag of high-speed train with pantographs is proposed. For the 10-car and 17-car train, the differences of total aerodynamic drag between the equation and the simulation results are 1.2% and 0.4%, respectively. The equation generalized in this study could well guide the design phase of high-speed train.
文摘本文充分利用网页数据的超链接关系和文本信息,提出了一种用于网页分类的归纳式半监督学习算法:基于图的Co-training网页分类算法(Graph based Co-training algorithmfor web page classification),简称GCo-training,并从理论上证明了算法的有效性.GCo-training在Co-training算法框架下,迭代地学习一个基于由超链接信息构造的图的半监督分类器和一个基于文本特征的Bayes分类器.基于图的半监督分类器只利用少量的标记数据,通过挖掘数据间大量的关系信息就可达到比较高的预测精度,可为Bayes分类器提供大量的标记信息;反过来学习大量标记信息后的Bayes分类器也可为基于图的分类器提供有效信息.迭代过程中,二者互相帮助,不断提高各自的性能,而后Bayes分类器可以用来预测大量未见数据的类别.在Web→KB数据集上的实验结果表明,与利用文本特征和锚文本特征的Co-training算法和基于EM的Bayes算法相比,GCo-training算法性能优越.
文摘Tri-training能有效利用无标记样例提高泛化能力.针对Tri-training迭代中无标记样例常被错误标记而形成训练集噪声,导致性能不稳定的缺点,文中提出ADE-Tri-training(Tri-training with Adaptive Data Editing)新算法.它不仅利用Remove Only剪辑操作对每次迭代可能产生的误标记样例识别并移除,更重要的是采用自适应策略来确定Remove Only触发与抑制的恰当时机.文中证明,PAC理论下自适应策略中一系列判别充分条件可同时确保新训练集规模迭代增大和新假设分类错误率迭代降低更多.UCI数据集上实验结果表明:ADE-Tri-training具有更好的分类泛化性能和健壮性.
