Softmax regression,which is also called multinomial logistic regression,is widely used in various fields for modeling the relationship between covariates and categorical responses with multiple levels.The increasing v...Softmax regression,which is also called multinomial logistic regression,is widely used in various fields for modeling the relationship between covariates and categorical responses with multiple levels.The increasing volumes of data bring new challenges for parameter estimation in softmax regression,and the optimal subsampling method is an effective way to solve them.However,optimal subsampling with replacement requires to access all the sampling probabilities simultaneously to draw a subsample,and the resultant subsample could contain duplicate observations.In this paper,the authors consider Poisson subsampling for its higher estimation accuracy and applicability in the scenario that the data exceed the memory limit.The authors derive the asymptotic properties of the general Poisson subsampling estimator and obtain optimal subsampling probabilities by minimizing the asymptotic variance-covariance matrix under both A-and L-optimality criteria.The optimal subsampling probabilities contain unknown quantities from the full dataset,so the authors suggest an approximately optimal Poisson subsampling algorithm which contains two sampling steps,with the first step as a pilot phase.The authors demonstrate the performance of our optimal Poisson subsampling algorithm through numerical simulations and real data examples.展开更多
A comprehensive assessment of the spatial-aware supervised learning algorithms for hyper-spectral image(HSI)classification was presented.For this purpose,standard support vector machines(SVMs),multinomial logistic reg...A comprehensive assessment of the spatial-aware supervised learning algorithms for hyper-spectral image(HSI)classification was presented.For this purpose,standard support vector machines(SVMs),multinomial logistic regression(MLR)and sparse representation(SR) based supervised learning algorithm were compared both theoretically and experimentally.Performance of the discussed techniques was evaluated in terms of overall accuracy,average accuracy,kappa statistic coefficients,and sparsity of the solutions.Execution time,the computational burden,and the capability of the methods were investigated by using probabilistic analysis.For validating the accuracy a classical benchmark AVIRIS Indian pines data set was used.Experiments show that integrating spectral-spatial context can further improve the accuracy,reduce the misclassification error although the cost of computational time will be increased.展开更多
Using disordered multinomial logistic regression and multiple linear regression method,385 copies of questionnaires on farmer are analyzed to explore the relationship between peasant's psychological traits,peasant...Using disordered multinomial logistic regression and multiple linear regression method,385 copies of questionnaires on farmer are analyzed to explore the relationship between peasant's psychological traits,peasant's cognition on seed technology and perception on supplydemand matching of new variety.Research results show that the vast majority of farmers think that current new variety is at high-level supplydemand balance and the oversupply status,and updating speed of new variety on the market is faster;the farmers preferring risk,seeking innovation and having strong learning and cognition ability may select high-level supply-demand matching state,and the farmers understanding the importance and difference of seed technology tend to choose high-level supply-demand matching situation;the farmers with strong learning and cognition ability can acknowledge the importance and difference of seed technology,while the farmers preferring risk can perceive the difference of seed technology;psychology seeking the innovation and learning and cognition ability affect the farmer's perception on supplydemand matching status of new variety via affecting the farmer's cognition on technical difference.展开更多
Although risk perception of natural hazards has been identified as an important determinant for sound policy design,there is limited empirical research on it in developing countries.This article narrows the empirical ...Although risk perception of natural hazards has been identified as an important determinant for sound policy design,there is limited empirical research on it in developing countries.This article narrows the empirical literature gap.It draws from Babessi,a rural town in the Northwest Region of Cameroon.Babessi was hit by a severe flash flood in 2012.The cross-disciplinary lens applied here deciphers the complexity arising from flood hazards,often embedded in contexts characterized by poverty,a state that is constrained in disaster relief,and market-based solutions being absent.Primary data were collected via snowball sampling.Multinomial logistic regression analysis suggests that individuals with leadership functions,for example,heads of households,perceive flood risk higher,probably due to their role as household providers.We found that risk perception is linked to location,which in turn is associated with religious affiliation.Christians perceive floods riskier than Muslims because the former traditionally reside at the foot of hills and the latter uphill;rendering Muslims less exposed and eventually less affected by floods.Finally,public disaster relief appears to have built up trust and subsequently reduced risk perception,even if some victims remained skeptical of state disaster relief.This indicates strong potential benefits of public transfers for flood risk management in developing countries.展开更多
基金Wang Haiying’s research was partially supported by the National Science Foundation under Grant No.CCF 2105571.
文摘Softmax regression,which is also called multinomial logistic regression,is widely used in various fields for modeling the relationship between covariates and categorical responses with multiple levels.The increasing volumes of data bring new challenges for parameter estimation in softmax regression,and the optimal subsampling method is an effective way to solve them.However,optimal subsampling with replacement requires to access all the sampling probabilities simultaneously to draw a subsample,and the resultant subsample could contain duplicate observations.In this paper,the authors consider Poisson subsampling for its higher estimation accuracy and applicability in the scenario that the data exceed the memory limit.The authors derive the asymptotic properties of the general Poisson subsampling estimator and obtain optimal subsampling probabilities by minimizing the asymptotic variance-covariance matrix under both A-and L-optimality criteria.The optimal subsampling probabilities contain unknown quantities from the full dataset,so the authors suggest an approximately optimal Poisson subsampling algorithm which contains two sampling steps,with the first step as a pilot phase.The authors demonstrate the performance of our optimal Poisson subsampling algorithm through numerical simulations and real data examples.
基金National Key Research and Development Program of China(No.2016YFF0103604)National Natural Science Foundations of China(Nos.61171165,11431015,61571230)+1 种基金National Scientific Equipment Developing Project of China(No.2012YQ050250)Natural Science Foundation of Jiangsu Province,China(No.BK20161500)
文摘A comprehensive assessment of the spatial-aware supervised learning algorithms for hyper-spectral image(HSI)classification was presented.For this purpose,standard support vector machines(SVMs),multinomial logistic regression(MLR)and sparse representation(SR) based supervised learning algorithm were compared both theoretically and experimentally.Performance of the discussed techniques was evaluated in terms of overall accuracy,average accuracy,kappa statistic coefficients,and sparsity of the solutions.Execution time,the computational burden,and the capability of the methods were investigated by using probabilistic analysis.For validating the accuracy a classical benchmark AVIRIS Indian pines data set was used.Experiments show that integrating spectral-spatial context can further improve the accuracy,reduce the misclassification error although the cost of computational time will be increased.
文摘Using disordered multinomial logistic regression and multiple linear regression method,385 copies of questionnaires on farmer are analyzed to explore the relationship between peasant's psychological traits,peasant's cognition on seed technology and perception on supplydemand matching of new variety.Research results show that the vast majority of farmers think that current new variety is at high-level supplydemand balance and the oversupply status,and updating speed of new variety on the market is faster;the farmers preferring risk,seeking innovation and having strong learning and cognition ability may select high-level supply-demand matching state,and the farmers understanding the importance and difference of seed technology tend to choose high-level supply-demand matching situation;the farmers with strong learning and cognition ability can acknowledge the importance and difference of seed technology,while the farmers preferring risk can perceive the difference of seed technology;psychology seeking the innovation and learning and cognition ability affect the farmer's perception on supplydemand matching status of new variety via affecting the farmer's cognition on technical difference.
文摘Although risk perception of natural hazards has been identified as an important determinant for sound policy design,there is limited empirical research on it in developing countries.This article narrows the empirical literature gap.It draws from Babessi,a rural town in the Northwest Region of Cameroon.Babessi was hit by a severe flash flood in 2012.The cross-disciplinary lens applied here deciphers the complexity arising from flood hazards,often embedded in contexts characterized by poverty,a state that is constrained in disaster relief,and market-based solutions being absent.Primary data were collected via snowball sampling.Multinomial logistic regression analysis suggests that individuals with leadership functions,for example,heads of households,perceive flood risk higher,probably due to their role as household providers.We found that risk perception is linked to location,which in turn is associated with religious affiliation.Christians perceive floods riskier than Muslims because the former traditionally reside at the foot of hills and the latter uphill;rendering Muslims less exposed and eventually less affected by floods.Finally,public disaster relief appears to have built up trust and subsequently reduced risk perception,even if some victims remained skeptical of state disaster relief.This indicates strong potential benefits of public transfers for flood risk management in developing countries.