This paper proposes a source localization solution robust to measurement outliers in time differences of arrivals(TDOA)measurements.The solution uses a piecewise loss function named as mixed Huber loss(MHL)proposed ba...This paper proposes a source localization solution robust to measurement outliers in time differences of arrivals(TDOA)measurements.The solution uses a piecewise loss function named as mixed Huber loss(MHL)proposed based on the classical Huber loss(HL)and its refined version.The MHL is able to effectively mitigate the impact of all levels of measurement outliers by setting two triggering thresholds.In practice,appropriate triggering threshold values can be obtained through simulation given the level of measurement noise and a rough range of potential measurement outliers.A clustering based approach is proposed to further improve the robustness of localization solution against reference sensor related outliers.Simulations are included to examine the solution's performance and compare it with several benchmarks.The proposed MHL based solution is shown to be superior to the classical solution and the benchmarks.The solution is shown to be even robust to multiple measurement outliers.Furthermore,the influence of range measurement outliers in the reference sensor can be effectively mitigated by the clustering based approach.展开更多
According to the World Health Organization,about 50 million people worldwide suffer from epilepsy.The detection and treatment of epilepsy face great challenges.Electroencephalogram(EEG)is a significant research object...According to the World Health Organization,about 50 million people worldwide suffer from epilepsy.The detection and treatment of epilepsy face great challenges.Electroencephalogram(EEG)is a significant research object widely used in diagnosis and treatment of epilepsy.In this paper,an adaptive feature learning model for EEG signals is proposed,which combines Huber loss function with adaptive weight penalty term.Firstly,each EEG signal is decomposed by intrinsic time-scale decomposition.Secondly,the statistical index values are calculated from the instantaneous amplitude and frequency of every component and fed into the proposed model.Finally,the discriminative features learned by the proposed model are used to detect seizures.Our main innovation is to consider a highly flexible penalization based on Huber loss function,which can set different weights according to the influence of different features on epilepsy detection.Besides,the new model can be solved by proximal alternating direction multiplier method,which can effectively ensure the convergence of the algorithm.The performance of the proposed method is evaluated on three public EEG datasets provided by the Bonn University,Childrens Hospital Boston-Massachusetts Institute of Technology,and Neurological and Sleep Center at Hauz Khas,New Delhi(New Delhi Epilepsy data).The recognition accuracy on these two datasets is 98%and 99.05%,respectively,indicating the application value of the new model.展开更多
For nonlinear state estimation driven by non-Gaussian noise,the estimator is required to be updated iteratively.Since the iterative update approximates a linear process,it fails to capture the nonlinearity of observat...For nonlinear state estimation driven by non-Gaussian noise,the estimator is required to be updated iteratively.Since the iterative update approximates a linear process,it fails to capture the nonlinearity of observation models,and this further degrades filtering accuracy and consistency.Given the flaws of nonlinear iteration,this work incorporates a recursive strategy into generalized M-estimation rather than the iterative strategy.The proposed algorithm extends nonlinear recursion to nonlinear systems using the statistical linear regression method.The recursion allows for the gradual release of observation information and consequently enables the update to proceed along the nonlinear direction.Considering the correlated state and observation noise induced by recursions,a separately reweighting strategy is adopted to build a robust nonlinear system.Analogous to the nonlinear recursion,a robust nonlinear recursive update strategy is proposed,where the associated covariances and the observation noise statistics are updated recursively to ensure the consistency of observation noise statistics,thereby completing the nonlinear solution of the robust system.Compared with the iterative update strategies under non-Gaussian observation noise,the recursive update strategy can facilitate the estimator to achieve higher filtering accuracy,stronger robustness,and better consistency.Therefore,the proposed strategy is more suitable for the robust nonlinear filtering framework.展开更多
The support vector machine(SVM)is a classical machine learning method.Both the hinge loss and least absolute shrinkage and selection operator(LASSO)penalty are usually used in traditional SVMs.However,the hinge loss i...The support vector machine(SVM)is a classical machine learning method.Both the hinge loss and least absolute shrinkage and selection operator(LASSO)penalty are usually used in traditional SVMs.However,the hinge loss is not differentiable,and the LASSO penalty does not have the Oracle property.In this paper,the huberized loss is combined with non-convex penalties to obtain a model that has the advantages of both the computational simplicity and the Oracle property,contributing to higher accuracy than traditional SVMs.It is experimentally demonstrated that the two non-convex huberized-SVM methods,smoothly clipped absolute deviation huberized-SVM(SCAD-HSVM)and minimax concave penalty huberized-SVM(MCP-HSVM),outperform the traditional SVM method in terms of the prediction accuracy and classifier performance.They are also superior in terms of variable selection,especially when there is a high linear correlation between the variables.When they are applied to the prediction of listed companies,the variables that can affect and predict financial distress are accurately filtered out.Among all the indicators,the indicators per share have the greatest influence while those of solvency have the weakest influence.Listed companies can assess the financial situation with the indicators screened by our algorithm and make an early warning of their possible financial distress in advance with higher precision.展开更多
基金supported by the National Natural Science Foundation of China(U20B2038)。
文摘This paper proposes a source localization solution robust to measurement outliers in time differences of arrivals(TDOA)measurements.The solution uses a piecewise loss function named as mixed Huber loss(MHL)proposed based on the classical Huber loss(HL)and its refined version.The MHL is able to effectively mitigate the impact of all levels of measurement outliers by setting two triggering thresholds.In practice,appropriate triggering threshold values can be obtained through simulation given the level of measurement noise and a rough range of potential measurement outliers.A clustering based approach is proposed to further improve the robustness of localization solution against reference sensor related outliers.Simulations are included to examine the solution's performance and compare it with several benchmarks.The proposed MHL based solution is shown to be superior to the classical solution and the benchmarks.The solution is shown to be even robust to multiple measurement outliers.Furthermore,the influence of range measurement outliers in the reference sensor can be effectively mitigated by the clustering based approach.
基金Supported by National Natural Science Foundation of China(Grant Nos.11701144,11971149)Henan Province Key and Promotion Special(Science and Technology)Project(Grant No.212102310305).
文摘According to the World Health Organization,about 50 million people worldwide suffer from epilepsy.The detection and treatment of epilepsy face great challenges.Electroencephalogram(EEG)is a significant research object widely used in diagnosis and treatment of epilepsy.In this paper,an adaptive feature learning model for EEG signals is proposed,which combines Huber loss function with adaptive weight penalty term.Firstly,each EEG signal is decomposed by intrinsic time-scale decomposition.Secondly,the statistical index values are calculated from the instantaneous amplitude and frequency of every component and fed into the proposed model.Finally,the discriminative features learned by the proposed model are used to detect seizures.Our main innovation is to consider a highly flexible penalization based on Huber loss function,which can set different weights according to the influence of different features on epilepsy detection.Besides,the new model can be solved by proximal alternating direction multiplier method,which can effectively ensure the convergence of the algorithm.The performance of the proposed method is evaluated on three public EEG datasets provided by the Bonn University,Childrens Hospital Boston-Massachusetts Institute of Technology,and Neurological and Sleep Center at Hauz Khas,New Delhi(New Delhi Epilepsy data).The recognition accuracy on these two datasets is 98%and 99.05%,respectively,indicating the application value of the new model.
基金co-supported by the National Natural Science Foundation of China(No.62303246,No.62103204)the China Postdoctoral Science Foundation(No.2023M731788)。
文摘For nonlinear state estimation driven by non-Gaussian noise,the estimator is required to be updated iteratively.Since the iterative update approximates a linear process,it fails to capture the nonlinearity of observation models,and this further degrades filtering accuracy and consistency.Given the flaws of nonlinear iteration,this work incorporates a recursive strategy into generalized M-estimation rather than the iterative strategy.The proposed algorithm extends nonlinear recursion to nonlinear systems using the statistical linear regression method.The recursion allows for the gradual release of observation information and consequently enables the update to proceed along the nonlinear direction.Considering the correlated state and observation noise induced by recursions,a separately reweighting strategy is adopted to build a robust nonlinear system.Analogous to the nonlinear recursion,a robust nonlinear recursive update strategy is proposed,where the associated covariances and the observation noise statistics are updated recursively to ensure the consistency of observation noise statistics,thereby completing the nonlinear solution of the robust system.Compared with the iterative update strategies under non-Gaussian observation noise,the recursive update strategy can facilitate the estimator to achieve higher filtering accuracy,stronger robustness,and better consistency.Therefore,the proposed strategy is more suitable for the robust nonlinear filtering framework.
文摘为了抑制采样点中粗差对数字高程模型(digital elevation model,DEM)建模的影响,以较高精度的多面函数(multi-quadric,MQ)为基函数,由改进Huber损失函数和权重惩罚项组成目标函数,发展了MQ抗差插值算法(MQ-H)。通过优化MQ-H目标函数,采样点权重计算最终转换为方程组求解。以数学曲面为研究对象,将MQ-H计算结果与传统MQ及最小绝对偏差MQ(MQ-L)进行比较,结果表明:当采样误差服从正态分布时,MQ-H计算精度与传统MQ相当,而远高于MQ-L;当采样误差服从拉普拉斯分布时,MQ-H计算精度略高于MQ-L及传统MQ;当采样点被粗差污染时,MQ-H计算精度远高于传统MQ及MQ-L。在实例分析中,以无人遥测飞艇立体像对获取的地面离散高程点为基础数据,基于MQ-H构建测区DEM,并将计算结果与传统插值算法,如反距离加权(inverse distance weighting,IDW)、普通克里金(ordinary Kriging,OK)和专业DEM插值软件ANUDEM(Australian National University DEM)进行比较,结果表明,传统插值方法在不同程度上受采样点中异常值或偶然误差影响,而MQ-H受异常值影响较小,且能准确捕捉到地形细节信息。
文摘The support vector machine(SVM)is a classical machine learning method.Both the hinge loss and least absolute shrinkage and selection operator(LASSO)penalty are usually used in traditional SVMs.However,the hinge loss is not differentiable,and the LASSO penalty does not have the Oracle property.In this paper,the huberized loss is combined with non-convex penalties to obtain a model that has the advantages of both the computational simplicity and the Oracle property,contributing to higher accuracy than traditional SVMs.It is experimentally demonstrated that the two non-convex huberized-SVM methods,smoothly clipped absolute deviation huberized-SVM(SCAD-HSVM)and minimax concave penalty huberized-SVM(MCP-HSVM),outperform the traditional SVM method in terms of the prediction accuracy and classifier performance.They are also superior in terms of variable selection,especially when there is a high linear correlation between the variables.When they are applied to the prediction of listed companies,the variables that can affect and predict financial distress are accurately filtered out.Among all the indicators,the indicators per share have the greatest influence while those of solvency have the weakest influence.Listed companies can assess the financial situation with the indicators screened by our algorithm and make an early warning of their possible financial distress in advance with higher precision.
