检索结果-内蒙古大学图书馆

Statistical inference methods and applications of outcome-dependent sampling designs under generalized linear models

引用

Science China Mathematics 2017年第7期60卷 1219-1238页

作者： YAN Shu DING JieLi LIU YanYan School of Mathematics and Statistics Wuhan University Wuhan 430072 China

A cost-effective sampling design is desirable in large cohort studies with a limited budget due to the high cost of measurements of primary exposure *** outcome-dependent sampling(ODS) designs enrich the observed sample by oversampling the regions of the underlying population that convey the most information about the exposure-response *** generalized linear models(GLMs) are widely used in many fields,however,much less developments have been done with the GLMs for data from the ODS *** study how to fit the GLMs to data obtained by the original ODS design and the two-phase ODS design,*** asymptotic properties of the proposed estimators are derived.A series of simulations are conducted to assess the finite-sample performance of the proposed *** to a Wilms tumor study and an air quality study demonstrate the practicability of the proposed methods.

关键词： biased-sampling two-phase design generalized linear models empirical likelihood

来源：评论

学校读者我要写书评

暂无评论

Robust Estimation of Self-Exciting generalized linear models With Application to Neuronal Modeling

引用

IEEE TRANSACTIONS ON SIGNAL PROCESSING 2017年第14期65卷 3733-3748页

作者： Kazemipour, Abbas Wu, Min Babadi, Behtash Univ Maryland Dept Elect & Comp Engn College Pk MD 20742 USA

We consider the problem of estimating self-exciting generalized linear models from limited binary observations, where the history of the process serves as the covariate. We analyze the performance of two classes of estimators, namely, the l(1)-regularized maximum likelihood and greedy estimators, for a canonical self-exciting process and characterize the sampling trade-offs required for stable recovery in the non-asymptotic regime. Our results extend those of compressed sensing for linear and generalized linear models with independent identically distributed covariates to those with highly interdependent covariates. We further provide simulation studies as well as application to real spiking data from the mouse's lateral geniculate nucleus and the ferret's retinal ganglion cells, which agree with our theoretical predictions.

关键词： Compressed sensing generalized linear models sparsity spontaneous activity neural signal processing

来源：评论

学校读者我要写书评

暂无评论

Instantaneous geometric rates via generalized linear models

引用

STATA JOURNAL 2017年第2期17卷 358-371页

作者： Discacciati, Andrea Bottai, Matteo Karolinska Inst Unit Biostat Stockholm Sweden Karolinska Inst Inst Environm Med Stockholm Sweden Karolinska Inst Biostat Unit Biostat Stockholm Sweden

The instantaneous geometric rate represents the instantaneous probability of an event of interest per unit of time. In this article, we propose a method to model the effect of covariates on the instantaneous geometric rate with two models: the proportional instantaneous geometric rate model and the proportional instantaneous geometric odds model. We show that these models can be fit within the generalized linear model framework by using two nonstandard link functions that we implement in the user-defined link programs log_igr and logit_igr. We illustrate how to fit these models and how to interpret the results with an example from a randomized clinical trial on survival in patients with metastatic renal carcinoma.

关键词： st0478 log_igr logit_igr instantaneous geometric rate generalized linear models glm survival analysis

来源：评论

学校读者我要写书评

暂无评论

Empirical likelihood for the response mean of generalized linear models with missing at random responses

引用

COMMUNICATIONS IN STATISTICS-SIMULATION AND COMPUTATION 2017年第1期46卷 164-173页

作者： Liu, Xiaohui Jiangxi Univ Finance & Econ Sch Stat Nanchang 330013 Jiangxi Peoples R China Jiangxi Univ Finance & Econ Res Ctr Appl Stat Nanchang Jiangxi Peoples R China

In this article, we present a new empirical likelihood ratio for constructing the confidence interval of the response mean of generalized linear models with missing at random responses. Compared with the existing methods, the proposal can avoid the so-called curse of dimensionality problem when the dimension of covariates is high, and is still chi-squared distributed asymptotically, nevertheless. Simulation studies are also provided to illustrate the performance of the developed method.

关键词： Empirical likelihood generalized linear models Missing at random Response mean

来源：评论

学校读者我要写书评

暂无评论

Predicting stem borer density in maize using RapidEye data and generalized linear models

引用

INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION 2017年 57卷 61-74页

作者： Abdel-Rahman, Elfatih M. Landmann, Tobias Kyalo, Richard Ong'amo, George Mwalusepo, Sizah Sulieman, Saad Le Ru, Bruno African Insect Sci Food & Hlth ICIPE POB 30772 Nairobi 00100 Kenya Univ Khartoum Dept Agron Fac Agr Khartoum 13314 Sudan Univ Nairobi Sch Biol Sci Coll Phys & Biol Sci Nairobi 00100 Kenya CNRS Lab Evolut Genomes Comportement & Ecol IRD CRNS UMR IRD EGCE 247 Gif Sur Yvette France Unversite Paris Sud 11 F-91405 Orsay France

Average maize yield in eastern Africa is 2.03 t ha(-1) as compared to gibbal average of 6.06 t ha(-1) due to biotic and abiotic constraints. Amongst the biotic production constraints in Africa, stem borers are the most injurious. In eastern Africa, maize yield losses due to stem borers are currently estimated between 12% and 21% of the total production. The objective of the present study was to explore the possibility of RapidEye spectral data to assess stem borer larva densities in maize fields in two study sites in Kenya. RapidEye images were acquired for the Bomet (western Kenya) test site on the 9th of December 2014 and on 27th of January 2015, and for Machakos (eastern Kenya) a RapidEye image was acquired on the 3rd of January 2015. Five RapidEye spectral bands as well as 30 specttal vegetation indices (SVIs) were utilized to predict per field maize stem borer larva densities using generalized linear models (GLMs), assuming Poisson ('Po') and negative binomial ('NB') distributions. Root mean square error (RMSE) and ratio prediction to deviation (RPD) statistics were used to assess the models performance using a leave one -out cross-validation approach. The Zero-inflated NB ('ZINB') models outperformed the 'NB' models and stem borer larva densities could only be predicted during the mid growing season in December and early January in both study sites, respectively (RMSE=0.69-1.06 and RPD = 8.25-19.57). Overall, all models performed similar when all the 30 SVIs (non-nested) and only the significant (nested) SVIs were used. The models developed could improve decision making regarding controlling maize stem borers within integrated pest management (IPM) interventions. (C) 2016 Elsevier B.V. All rights reserved.

关键词： RapidEye Maize Stem borer density generalized linear models Africa Crop productivity

来源：评论

学校读者我要写书评

暂无评论

Relationship between neighbourhood socioeconomic position and neighbourhood public green space availability: An environmental inequality analysis in a large German city applying generalized linear models

引用

INTERNATIONAL JOURNAL OF HYGIENE AND ENVIRONMENTAL HEALTH 2017年第4期220卷 711-718页

作者： Schuele, Steffen Andreas Gabriel, Katharina M. A. Bolte, Gabriele Univ Bremen Inst Publ Hlth & Nursing Res Dept Social Epidemiol Bremen Germany Univ Bremen Hlth Sci Bremen Bremen Germany

Background: The environmental justice framework states that besides environmental burdens also resources may be social unequally distributed both on the individual and on the neighbourhood level. This ecological study investigated whether neighbourhood socioeconomic position (SEP) was associated with neighbourhood public green space availability in a large German city with more than 1 million inhabitants. Methods: Two different measures were defined for green space availability. Firstly, percentage of green space within neighbourhoods was calculated with the additional consideration of various buffers around the boundaries. Secondly, percentage of green space was calculated based on various radii around the neighbourhood centroid. An index of neighbourhood SEP was calculated with principal component analysis. Log-gamma regression from the group of generalized linear models was applied in order to consider the non-normal distribution of the response variable. All models were adjusted for population density. Results: Low neighbourhood SEP was associated with decreasing neighbourhood green space availability including 200 m up to 1000 m buffers around the neighbourhood boundaries. Low neighbourhood SEP was also associated with decreasing green space availability based on catchment areas measured from neighbourhood centroids with different radii (1000m up to 3000 m). With an increasing radius the strength of the associations decreased. Conclusions: Social unequally distributed green space may amplify environmental health inequalities in an urban context. Thus, the identification of vulnerable neighbourhoods and population groups plays an important role for epidemiological research and healthy city planning. As a methodical aspect, log-gamma regression offers an adequate parametric modelling strategy for positively distributed environmental variables. (C) 2017 Elsevier GmbH. All rights reserved.

关键词： Green space Neighbourhood Environmental justice generalized linear models Health inequalities Environmental inequalities

来源：评论

学校读者我要写书评

暂无评论

Inference for generalized linear models via Alternating Directions and Bethe Free Energy Minimization

引用

IEEE TRANSACTIONS ON INFORMATION THEORY 2017年第1期63卷 676-697页

作者： Rangan, Sundeep Fletcher, Alyson K. Schniter, Philip Kamilov, Ulugbek S. NYU Dept Elect & Comp Engn Brooklyn NY 11201 USA Univ Calif Los Angeles Dept Stat Los Angeles CA 90095 USA Univ Calif Los Angeles Dept Math Los Angeles CA 90095 USA Univ Calif Los Angeles Dept Elect Engn Los Angeles CA 90095 USA Ohio State Univ Dept Elect & Comp Engn Columbus OH 43210 USA Ecole Polytech Fed Lausanne Biomed Imaging Grp CH-1015 Lausanne Switzerland Mitsubishi Elect Res Labs Cambridge MA 02139 USA

generalized linear models, where a random vector x is observed through a noisy, possibly nonlinear, function of a linear transform z = Ax, arise in a range of applications in nonlinear filtering and regression. Approximate message passing (AMP) methods, based on loopy belief propagation, are a promising class of approaches for approximate inference in these models. AMP methods are computationally simple, general, and admit precise analyses with testable conditions for optimality for large i.i.d. transforms A. However, the algorithms can diverge for general A. This paper presents a convergent approach to the generalized AMP (GAMP) algorithm based on direct minimization of a large-system limit approximation of the Bethe free energy (LSL-BFE). The proposed method uses a double-loop procedure, where the outer loop successively linearizes the LSL-BFE and the inner loop minimizes the linearized LSL-BFE using the alternating direction method of multipliers (ADMM). The proposed method, called ADMM-GAMP, is similar in structure to the original GAMP method, but with an additional least-squares minimization. It is shown that for strictly convex, smooth penalties, ADMM-GAMP is guaranteed to converge to a local minimum of the LSL-BFE, thus providing a convergent alternative to GAMP that is stable under arbitrary transforms. Simulations are also presented that demonstrate the robustness of the method for non-convex penalties as well.

关键词： Belief propagation ADMM message passing variational optimization generalized linear models

来源：评论

学校读者我要写书评

暂无评论

Inference for generalized linear models via Alternating Directions and Bethe Free Energy Minimization

Inference for Generalized Linear Models via Alternating Dire...

引用

IEEE International Symposium on Information Theory (ISIT)

ISBN: (纸本)9781467377041

关键词： Belief propagation ADMM message passing variational optimization generalized linear models

来源：评论

学校读者我要写书评

暂无评论

l₁ regularized multiplicative iterative path algorithm for non-negative generalized linear models

引用

COMPUTATIONAL STATISTICS & DATA ANALYSIS 2016年 101卷 289-299页

作者： Mandal, B. N. Ma, Jun ICAR Indian Agr Stat Res Inst New Delhi 110012 India Macquarie Univ Dept Stat N Ryde NSW 2109 Australia

In regression modeling, often a restriction that regression coefficients are non-negative is faced. The problem of model selection in non-negative generalized linear models (NNGLM) is considered using lasso, where regression coefficients in the linear predictor are subject to non-negative constraints. Thus, non-negatively constrained regression coefficient estimation is sought by maximizing the penalized likelihood (such as the l(1)-norm penalty). An efficient regularization path algorithm is proposed for generalized linear models with non-negative regression coefficients. The algorithm uses multiplicative updates which are fast and simultaneous. Asymptotic results are also developed for the constrained penalized likelihood estimates. Performance of the proposed algorithm is shown in terms of computational time, accuracy of solutions and accuracy of asymptotic standard deviations. (C) 2016 Elsevier B.V. All rights reserved.

关键词： generalized linear models Lasso Elastic net l(1)-norm penalty Regularization path Non-negativity constraints

来源：评论

学校读者我要写书评

暂无评论

Minimax designs for 2^k factorial experiments for generalized linear models

引用

COMMUNICATIONS IN STATISTICS-THEORY AND METHODS 2016年第16期45卷 4788-4797页

作者： Normark, Sofia Stockholm Univ Dept Stat SE-10691 Stockholm Sweden

Formulas for A- and C-optimal allocations for binary factorial experiments in the context of generalized linear models are derived. Since the optimal allocations depend on GLM weights, which often are unknown, a minimax strategy is considered. This is shown to be simple to apply to factorial experiments. Efficiency is used to evaluate the resulting design. In some cases, the minimax design equals the optimal design. For other cases no general conclusion can be drawn. An example of a two-factor logit model suggests that the minimax design performs well, and often better than a uniform allocation.

关键词： A-optimality C-optimality Factorial designs generalized linear models Minimax designs

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：