dglm.control: Auxiliary for Controlling dglm Fitting

In glm.ddR: Distributed 'glm' for Big Data using 'ddR' API

Description

Auxiliary function for dglm fitting. Typically only used internally by dglm.fit, but may be used to construct a control argument to either function.

Usage

dglm.control(epsilon = 1e-08, maxit = 25, trace = FALSE, 
               rigorous = FALSE)

Arguments

epsilon	It is used to adjust desired accuracy of the result.
maxit	It is the maximu, number of iterations before achieving the desired accuracy.
trace	When this argument is true, intermediate steps of the progress are displayed.
rigorous	When this argument is true, some extra checks are performed during fitting procedure. For example, mu and eta may be validating in each iteration to check if the fitted values are outside of the domain. Usually these checks are time consuming; therefore, the default value for this argument is FALSE.

Value

A list with components named as the arguments.

Examples

 ## Not run: 
    library(glm.ddR)

    Y <- as.darray(as.matrix(mtcars$am),c(ceiling(length(mtcars$am)/4),1))
    X <- as.darray(as.matrix(cbind(mtcars$wt,mtcars$hp)),
                   c(ceiling(length(mtcars$hp)/4),2))

    myModel <- dglm(Y, X, binomial, control=list(epsilon=1e-02, maxit=5,
                      trace=FALSE, rigorous=TRUE))
 
## End(Not run)

glm.ddR documentation built on May 29, 2017, 6:49 p.m.