检索结果-内蒙古大学图书馆

作者： Ficek, Joseph University of South Florida

学位级别：Ph.D., Doctor of Philosophy

Background: There is a need for rigorous and standardized methods of privacy protection for shared data in the health sciences. Differential privacy is one such method that has gained much popularity due to its versatility and robustness. This study evaluates differential privacy for explanatory regression modeling in the context of health research. Methods: Surveyed and newly proposed algorithms were evaluated with respect to the accuracy (bias and RMSE) of coefficient estimates, the empirical coverage probability of confidence intervals, and the power and type I error rates of hypothesis tests. Evaluations took place in both simulated and real data from a study of adolescent behavioral health. Results: For coefficient estimation, the simulation found the objective and output perturbation algorithms to be the most accurate for logistic models, and subsample-and-aggregate emerged as the most accurate for linear and log-linear models. However, only objective and output perturbation had sufficiently low noise at reasonable settings of the privacy parameter epsilon. The empirical coverage probability of confidence intervals only neared the nominal 95% rate for the output perturbation algorithm, at less private settings of epsilon. Of the available algorithms for hypothesis testing, only the Noisy Aggregated Censored z-test maintained an appropriate type I error rate, though power was only satisfactory at the least private settings of epsilon. Conclusions: The objective and output perturbation algorithms emerged as the most promising for differentially private regression statistics. Further work is needed to derive corresponding algorithms for statistical inference.

关键词： Confidentiality Data privacy generalized linear models Hypothesis testing Statistical disclosure limitation

来源：评论

学校读者我要写书评

暂无评论

Adoption of crop insurance in Ghana: an application of the complementary log-log truncated Poisson double-hurdle model

引用

AGRICULTURAL FINANCE REVIEW 2021年第1期81卷 76-93页

作者： Addey, Kwame Asiam Jatoe, John Baptist D. Kwadzo, George Tsey-Mensah North Dakota State Univ Ctr Agr Policy & Trade Studies Fargo ND USA Univ Ghana Dept Agr Econ & Agribusiness Accra Ghana Univ Ghana Accra Ghana

Purpose The aim of this paper is to identify the factors that influence rice farmers' decisions to adopt crop insurance and premium payments (willingness to pay [WTP] amounts). The paper also demonstrates the usefulness of the complementary log-log (cloglog) truncated Poisson double-hurdle model as an alternative hurdle model. Design/methodology/approach The study first investigated the nature of the dependent variable, which had non-normal residuals and was overdispersed. The probit truncated normal regression double-hurdle model was tried but it failed the normality and homoscedasticity tests;hence, the cloglog truncated Poisson double-hurdle model was employed in the study. Findings An estimated 61% of respondents would purchase crop insurance, despite farmers not having prior experience with this product. Amongst others, the factors that influence insurance adoption amongst rice farmers are the share of rice in total income, reliability perception of crop insurance schemes and the probability of failure to achieve target yields. The latter helps the authors to address adverse selection, a central issue to the viability of such an insurance programme. The determinants of farmers' WTP are also identified. Research limitations/implications Sampling was limited to farmers using irrigation and living in one region of Ghana, which may limit the study's wider applicability. Originality/value As far as the authors are aware, this study is the first to select the appropriate hurdle model based on established properties of the dependent variable on this topic - crop insurance decisions.

关键词： Crop insurance Cloglog truncated Poisson double-hurdle model Moments-equality Rice generalized linear models Fano factor Ghana

来源：评论

学校读者我要写书评

暂无评论

On robust estimation of negative binomial INARCH models

引用

METRON-INTERNATIONAL JOURNAL OF STATISTICS 2021年第2期79卷 137-158页

作者： Elsaied, Hanan Fried, Roland Suez Canal Univ Fac Commerce Dept Stat Ismailia 41522 Egypt TU Dortmund Univ Dept Stat D-44221 Dortmund Germany

We discuss robust estimation of INARCH models for count time series, where each observation conditionally on its past follows a negative binomial distribution with a constant scale parameter, and the conditional mean depends linearly on previous observations. We develop several robust estimators, some of them being computationally fast modifications of methods of moments, and some rather efficient modifications of conditional maximum likelihood. These estimators are compared to related recent proposals using simulations. The usefulness of the proposed methods is illustrated by a real data example.

关键词： Count time series Negative binomial distribution Overdispersion generalized linear models Rank autocorrelation Tukey M-estimator Additive outliers

来源：评论

学校读者我要写书评

暂无评论

Semiparametric Bayesian inference on generalized linear measurement error models

引用

STATISTICAL PAPERS 2017年第4期58卷 1091-1113页

作者： Tang, Nian-Sheng Li, De-Wang Tang, An-Min Yunnan Univ Dept Stat Kunming 650091 Yunnan Peoples R China

The classical assumption in generalized linear measurement error models (GLMEMs) is that measurement errors (MEs) for covariates are distributed as a fully parametric distribution such as the multivariate normal distribution. This paper uses a centered Dirichlet process mixture model to relax the fully parametric distributional assumption of MEs, and develops a semiparametric Bayesian approach to simultaneously obtain Bayesian estimations of parameters and covariates subject to MEs by combining the stick-breaking prior and the Gibbs sampler together with the Metropolis-Hastings algorithm. Two Bayesian case-deletion diagnostics are proposed to identify influential observations in GLMEMs via the Kullback-Leibler divergence and Cook's distance. Computationally feasible formulae for evaluating Bayesian case-deletion diagnostics are presented. Several simulation studies and a real example are used to illustrate our proposed methodologies.

关键词： Cook's distance Dirichlet process prior generalized linear models Kullback-Leibler divergence Measurement error models

来源：评论

学校读者我要写书评

暂无评论

Non-Stationary Influences of Large-Scale Climate Drivers on Low Flow Extremes in Southeast Australia

引用

WATER RESOURCES RESEARCH 2022年第7期58卷 e2021WR031508-e2021WR031508页

作者： Goswami, Pallavi Peterson, Tim J. Mondal, Arpita Rudiger, Christoph Indian Inst Technol IITB Monash Res Acad Mumbai Maharashtra India Monash Univ Dept Civil Engn Clayton Vic Australia Univ Melbourne Dept Infrastruct Engn Parkville Vic Australia Indian Inst Technol Dept Civil Engn Mumbai Maharashtra India Indian Inst Technol Interdisciplinary Program Climate Studies Mumbai Maharashtra India Bur Meteorol Melbourne Vic Australia

The intensity, duration, and frequency (IDF) of low streamflow events are often considered stationary in time. Given droughts are likely to become more frequent and extreme, here those IDFs are analyzed for their non-stationarity in relation to three climate drivers, namely, the El Nino-Southern Oscillation, the Indian Ocean Dipole, and the Southern Annular Mode. For this, low flow events were identified for 161 unregulated catchments in Victoria, Australia. The relationship between the climate drivers and the IDFs were developed using covariate-based non-stationary statistical models. We found that several of these catchments exhibit strong non-stationary influences from the climate modes. Such catchments have a higher probability of experiencing more intense, prolonged, and/or frequent low flows when a dry phase of these climate drivers occurs. We found that catchments having strong evidence of non-stationary frequency (n = 77) were more compared to those having non-stationary intensity (n = 19) or duration (n = 34). Further, we derived projected indices of these climate drivers from CMIP5 models for the RCP8.5 scenario. These indicate an intensification in the dry phase of the climate drivers. Using information of the projected indices in the IDF models, we found that future dry phases of these climate drivers are likely to further increase the probability of experiencing higher magnitudes of low flow IDFs in catchments non-stationary in these characteristics in relation to one or more of these climate drivers. Frequency shows greater response to future dry phases of these drivers than intensity or duration of low flows.

关键词： IDF of low flows non-stationarity droughts climate teleconnections extreme value theory generalized linear models

来源：评论

学校读者我要写书评

暂无评论

Statistical Attribution of the Influence of Urban and Tree Cover Change on Streamflow: A Comparison of Large Sample Statistical Approaches

引用

WATER RESOURCES RESEARCH 2022年第5期58卷 e2021WR030742-e2021WR030742页

作者： Anderson, Bailey J. Slater, Louise J. Dadson, Simon J. Blum, Annalise G. Prosdocimi, Ilaria Univ Oxford Sch Geog & Environm Oxford England NERC Ctr Ecol & Hydrol Wallingford Oxon England Johns Hopkins Univ Washington DC USA Univ Ca Foscari Venezia Dipartimento Sci Ambientali Informat & Stat Venice Italy

The strengths and weaknesses of different statistical methodologies for attributing changes in streamflow to land cover are still poorly understood. We examine the relationships between high (Q(99)), mean (Q(mean)), and low (Q(01)) streamflow and urbanization or tree cover change in 729 catchments in the United States between 1992 and 2018. We apply two statistical modeling approaches and compare their performance. Panel regression models estimate the average effect of land cover changes on streamflow across all sites, and show that on average, a 1%-point increase in catchment urban area results in a small (0.6%-0.7%), but highly significant increase in mean and high flows. Meanwhile, a 1%-point increase in tree cover does not correspond to strongly significant changes in flow. We also fit a generalized linear model to each individual site, which results in highly varied model coefficients. The medians of the single-site coefficients show no significant relationships between either urbanization or tree cover change and any streamflow quantile (although at individual sites, the coefficients may be statistically significant and positive or negative). On the other hand, the GLM coefficients may provide greater nuance in catchments with specific attributes. This variation is not well represented through the panel model estimates of average effect, unless moderators are carefully considered. We highlight the value of statistical approaches for large-sample attribution of hydrological change, while cautioning that considerable variability exists.

关键词： statistical hydrology large-sample panel regression land cover change generalized linear models

来源：评论

学校读者我要写书评

暂无评论

Implicit differentiation for fast hyperparameter selection in non-smooth convex learning

The Journal of Machine Learning Research

引用

The Journal of Machine Learning Research 2022年第1期23卷 6680-6722页

作者： Quentin Bertrand Quentin Klopfenstein Mathurin Massias Mathieu Blondel Samuel Vaiter Alexandre Gramfort Joseph Salmon Université Paris-Saclay Inria CEA Palaiseau France Institut Mathématique de Bourgogne Université de Bourgogne Dijon France MaLGa DIBRIS Università degli Studi di Genova Genova Italy Google Research Brain team Paris France CNRS and Institut Mathématique de Bourgogne Université de Bourgogne Dijon France IMAG Université de Montpellier CNRS Montpellier France

Finding the optimal hyperparameters of a model can be cast as a bilevel optimization problem, typically solved using zero-order techniques. In this work we study first-order methods when the inner optimization problem is convex but non-smooth. We show that the forward-mode differentiation of proximal gradient descent and proximal coordinate descent yield sequences of Jacobians converging toward the exact Jacobian. Using implicit differentiation, we show it is possible to leverage the non-smoothness of the inner problem to speed up the computation. Finally, we provide a bound on the error made on the hypergradient when the inner optimization problem is solved approximately. Results on regression and classification problems reveal computational benefits for hyperparameter optimization, especially when multiple hyperparameters are required.

关键词： convex optimization hyperparameter optimization hyperparameter selection bilevel optimization Lasso generalized linear models

来源：评论

学校读者我要写书评

暂无评论

Variation in abundance and habitat relationship of three understory insectivorous birds in a disturbed landscape of Neotropical dry forest of Colombia

引用

Avian Research 2020年第3期11卷 265-276页

作者： Camilo Loaiza Olga LMontenegro David King Manuel Spínola Lilibeth Arinda Palacio Agustín Rudas Instituto de Ciencias Naturales Universidad Nacional de ColombiaBogotáColombia Northern Research Station USDA Forest ServiceUniversity of Massachusetts204 Holdsworth HallAmherstMA 01003USA Instituto Internacional en Conservación y Manejo de Vida Silvestre Universidad NacionalApartado1350-3000 HerediaCosta Rica.

Background:Evaluating relationships between avian populations and their habitat is important for understanding the biology of these species and for management decisions that improve the effectiveness of restoration *** investigated how habitat variation along a disturbance gradient affected the abundance of three insectivorous bird species,White-bellied Antbird(Myrmeciza longipes),Barred Antshrike(Thamnophilus doliatus),and Pale-breasted Spinetail(Synallaxis albescens)in a Neotropical dry forest remnant in central ***:The study area,which is located in Magdalena Valley,is under ecological restoration after the construction of the“El Quimbo”hydroelectric dam,with a gradient of habitat conditions,from relatively advanced secondary growth to recently disturbed *** conducted bird surveys and quantified the habitat conditions from June 2018 to June *** compared the relative abundance of the birds among the three zones with increasing disturbance and modeled the relationship between the bird abundance with the micro-and landscape-habitat conditions,with Poisson generalized linear model ***:The relative abundance of the three bird species varied according to the habitat features and disturbance *** abundance of White-bellied Antbirds was higher in forests and thickets than shrubland fragments;Barred Antshrike was more abundant in shrubland and thickets than in forest fragments,and Pale-breasted Spinetail had the highest abundance in shrubland and the lowest in forest *** abundance of both White-bellied Antbird and Pale-breasted Spinetail did not differ among zones with different levels of disturbance,while the abundance of Barred Antshrike decreased in zones with more ***:There were significant differences in the habitat relationships of the three understory insectivorous species along the disturbance gradient,although these species were within the same trophic *** bird species prefer

关键词： Bird population ecology generalized linear models Wildlife habitat restoration

来源：评论

学校读者我要写书评

暂无评论

Robust linear Regression and Anomaly Detection in the Presence of Poisson Noise Using Expectation-Propagation 2nd

Robust Linear Regression and Anomaly Detection in the Presen...

引用

2nd World Congress on Condition Monitoring, WCCM 2019

作者： Altmann, Yoann Yao, Dan McLaughlin, Stephen Davies, Mike E. School of Engineering and Physical Science Heriot-Watt University Edinburgh United Kingdom School of Engineering University of Edinburgh Edinburgh United Kingdom

ISBN: (纸本)9789811591983

This paper presents a family of approximate Bayesian methods for joint anomaly detection and linear regression in the presence of non-Gaussian noise. Robust anomaly detection using non-convex sparsity-promoting regularization terms is generally challenging, in particular when additional uncertainty measures about the estimation process are needed, e.g., posterior probabilities of anomaly presence. The problem becomes even more challenging in the presence of non-Gaussian, (e.g., Poisson distributed), additional constraints on the regression coefficients (e.g., positivity) and when the anomalies present complex structures (e.g., structured sparsity). Uncertainty quantification is classically addressed using Bayesian methods. Specifically, Monte Carlo methods are the preferred tools to handle complex models. Unfortunately, such simulation methods suffer from a significant computational cost and are thus not scalable for fast inference in high dimensional problems. In this paper, we thus propose fast alternatives based on Expectation-Propagation (EP) methods, which aim at approximating complex distributions by more tractable models to simplify the inference process. The main problem addressed in this paper is linear regression and (sparse) anomaly detection in the presence of noisy measurements. The aim of this paper is to demonstrate the potential benefits and assess the performance of such EP-based methods. The results obtained illustrate that approximate methods can provide satisfactory results with a reasonable computational cost. It is important to note that the proposed methods are sufficiently generic to be used in other applications involving condition monitoring. © 2021, Springer Nature Singapore Pte Ltd.

关键词： Anomaly detection Bayesian inference Expectation-propagation generalized linear models Robust regression

来源：评论

学校读者我要写书评

暂无评论

Instrumental variable approach to covariate measurement error in generalized linear models

引用

ANNALS OF THE INSTITUTE OF STATISTICAL MATHEMATICS 2012年第3期64卷 475-493页

作者： Abarin, Taraneh Wang, Liqun Univ Manitoba Dept Stat Winnipeg MB R3T 2N2 Canada

The paper presents the method of moments estimation for generalized linear measurement error models using the instrumental variable approach. The measurement error has a parametric distribution that is not necessarily normal, while the distributions of the unobserved covariates are nonparametric. We also propose simulation-based estimators for the situation where the closed forms of the moments are not available. The proposed estimators are strongly consistent and asymptotically normally distributed under some regularity conditions. Finite sample performances of the estimators are investigated through simulation studies.

关键词： Errors in variables generalized linear models Heterogeneity Measurement error Instrumental variable Method of moments M-estimation Nonlinear models Simulation-based estimation

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：