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 pa
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.
fits a generalized linear model E[y|x] = G(x*b). This is the core routine 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.
glmest fits a generalized linear model E[y|x] = G(x*b). This routine offers a convenient interface for GLM estimation. A check of the data is performed.
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.
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 va
glmopt defines a list with optional parameters in glm functions. 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.
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.
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.