Titlebook: Vector Generalized Linear and Additive Models; With an Implementati Thomas W. Yee Textbook 2015 Thomas Yee 2015 Additive Model Theory.Gener

遗产

Univariate Continuous Distributionslated according to their support, and/or the distribution from which they are derived (e.g., beta-type, gamma-type), and/or their purpose (e.g., statistical size distributions, actuarial distributions).

defuse

Categorical Data Analysisl logit model for nominal responses; and for ordinal responses: the proportional and non-proportional-odds models, continuation and stopping ratio models, and the adjacent categories model. Some other topics includes the . argument for allowing ..-specific covariates, the Poisson trick, marginal effects, and genetic models.

积习已深

Vector Generalized Linear and Additive Models978-1-4939-2818-7Series ISSN 0172-7397 Series E-ISSN 2197-568X

渗入

VGLMsThis chapter covers the most important class of models, viz. vector generalized linear models (VGLMs). It includes the basic ideas of (parameter) link functions, constraint matrices, the . argument, inference, computational details, (e.g., QR decomposition, Cholesky), residuals and diagnostics. Some basic details on software usage is given.

赞美者

Other TopicsThis chapter looks at a selection of miscellaneous topics, such as calculation of working weights by simulated Fisher scoring (SFS), information criteria for model selection, and bias-reduction (for GLMs). The latter has been used to obtain a finite solution to completely separated binary data.

合同

Bivariate Continuous DistributionsThis chapter lists a small number of bivariate distributions whose parameters can easily be estimated by IRLS. A handful of a special type of bivariate distribution, called copulas, are also implemented. Some special consideration is given to the bivariate normal distribution and Plackett’s bivariate distribution.

emulsify

旋转一周

Thomas W. YeeFirst comprehensive book on Vector Generalized Linear Models.Builds on success of VGAM R package, which is used to show applications of the methodology.Techniques on regression and parametric and non-