Keywords - Function groups - @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Library: glm

doglm doglm provides an interactive menu for preparing data and fitting GLM in XploRe.
genbilo genbilo generates data from a logit model, i.e. y from Binomial(m,p) with P(y=1)=1./(1+exp(-x*b))
genmultlo genmultlo generates data according to a multinomial logit model with P( Y = j | Xa , Xi) proportional to exp( Xa * ba + Xi * bi[j] ). Here, Xi denotes the part of the explanatory variables which merely depends on the individuals, Xa covers variables which may vary with the alternatives j. Either part, Xa or Xi, can be omitted.
glmbackward glmbackward performs a backward model selection by searching the best of all subset models w.r.t. the AIC or BIC criterion. Optionally, a number of columns can be given, which are always included in the submodels.
glmbilo glmbilo fits a generalized linear model where y|x is binomial distributed and E[y|x] and x*b are linked via the logistic function (canonical link)
glmbipro glmbipro fits a generalized linear model where y|x is binomial distributed and E[y|x] and x*beta are linked via the gaussian cdf (non canonical link)
glmcore fits a generalized linear model E[y|x] = G(x*b). This is the core macro for GLM estimation. It assumes that all input variables are given in the right manner. No preparation of data is performed. A more convenient way to estimate a GLM is to call the function glmest.
glmdiagh glmdiagh calculates the diagonal elements of the 'hat' matrix
glmest glmest fits a generalized linear model E[y|x] = G(x*b). This macro offers a convenient interface for GLM estimation. A check of the data is performed.
glmfit helper macro for doglm.
glmforward glmforward performs a forward model selection by searching the best of all subset models w.r.t. the AIC or BIC criterion. Optionally, a number of columns can be given, which are always included in the submodels.
glminit glminit checks the validity of input and performs the initial calculations for an GLM fit. The output is ready to be used with glmcore.
glminvlink glminvlink computes the inverse link function.
glmlink glmlink computes the link function.
glmll glmll computes the individual log-likelihood.
glmlld glmlld computes the first and second derivative of the individual log-likelihood in dependence of the linear index eta and y.
glmlrtest glmlrtest performs a likelihood ratio test of two nested GLM.
glmmain sets defaults for library glm.
glmmultlo glmmultlo fits a multinomial/conditional logit model where the response Y is multinomial distributed. This means, P( Y = j | Xa , Xi) is proportional to exp( Xa * ba + Xi * bi[j] ). Here, Xi denotes that part of the explanatory variables which merely depends on the individuals and Xa covers variables which may vary with the alternatives j. Either part, Xa or Xi, can be omitted.
glmmultshape reshapes data from panel format to matrix format and vice versa. The matrix format is needed for glmmultlo.
glmnoid glmnoid fits a generalized linear model where y|x is normally distributed and E[y|x] and x*beta are linked via the identity function (canonical link)
glmopt glmopt defines a list with optional parameters in glm macros. The list is either created or new options are appended to an existing list. Note that glmopt does accept _any_ values for the parameters without validity.
glmout glmout creates a nice output display for GLM.
glmplot glmplot creates a display and plots for a one-dimensional explanatory variable: the distribution, a scatterplot of the marginal influence versus the response and a scatterplot of the variabel versus the response.
glmscatter glmscatter computes a scatterplot to explain the marginal influence of a variable on the response.
glmselect glmselect performs a model selection by searching the best of all subset models w.r.t. the AIC or BIC criterion. Optionally a number of columns can be given, which are always included in the submodels.
glmstat glmstat provides some statistics for a fitted GLM.
glmtest executes some tests for the macros defined in glm.lib. Is invoked by vertestl().
select0item select0item is an item selector for glm. It allows exactly at most one item to be selected.
select1item select1item is an item selector for glm. It allows exactly one item to be selected.
xlnxdy computes x.*log(x./y), with check for x=y=0, used for GLM macros.
xlny computes x.*log(y), with check for x=y=0, used for GLM macros.

Keywords - Function groups - @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

(C) MD*TECH Method and Data Technologies, 28.6.1999