GLM_Exact_Design: rounding algorithm for generalized linear models

In ForLion: 'ForLion' Algorithms to Find Optimal Experimental Designs with Mixed Factors

rounding algorithm for generalized linear models

Description

rounding algorithm for generalized linear models

Usage

GLM_Exact_Design(
  k.continuous,
  design_x,
  design_p,
  det.design,
  p,
  ForLion,
  bvec,
  Integral_based,
  b_matrix,
  joint_Func_b,
  Lowerbounds,
  Upperbounds,
  rel.diff,
  L,
  N,
  hfunc,
  link
)

Arguments

k.continuous	number of continuous factors
design_x	the matrix with rows indicating design point which we got from the approximate design
design_p	the corresponding approximate allocation
det.design	the determinant of the approximate design
p	number of parameters
ForLion	TRUE or FALSE, TRUE: this approximate design was generated by ForLion algorithm, FALSE: this approximate was generated by EW ForLion algorithm
bvec	If ForLion==TRUE assumed parameter values of model parameters beta, same length of h(y)
Integral_based	TRUE or FALSE, if TRUE then we will find the integral-based EW D-optimality otherwise we will find the sample-based EW D-optimality
b_matrix	The matrix of the sampled parameter values of beta
joint_Func_b	The prior joint probability distribution of the parameters
Lowerbounds	The lower limit of the prior distribution for each parameter
Upperbounds	The upper limit of the prior distribution for each parameter
rel.diff	points with distance less than that will be merged
L	vector: rounding factors
N	total number of observations
hfunc	function for obtaining model matrix h(y) for given design point y, y has to follow the same order as n.factor
link	link function, default "logit", other links: "probit", "cloglog", "loglog", "cauchit", "log", "identity"

Value

x.design matrix with rows indicating design point

ni.design exact allocation

rel.efficiency relative efficiency of the Exact and Approximate Designs

Examples

k.continuous=1
design_x=matrix(c(25, -1, -1,-1, -1 ,
                 25, -1, -1, -1, 1,
                 25, -1, -1, 1, -1,
                 25, -1, -1, 1, 1,
                 25, -1, 1, -1, -1,
                 25, -1, 1, -1, 1,
                 25, -1, 1, 1, -1,
                 25, -1, 1, 1, 1,
                 25, 1, -1, 1, -1,
                 25, 1, 1, -1, -1,
                 25, 1, 1, -1, 1,
                 25, 1, 1, 1, -1,
                 25, 1, 1, 1, 1,
                 38.9479, -1, 1, 1, -1,
                 34.0229, -1, 1, -1, -1,
                 35.4049, -1, 1, -1, 1,
                 37.1960, -1, -1, 1, -1,
                 33.0884, -1, 1, 1, 1),nrow=18,ncol=5,byrow = TRUE)
hfunc.temp = function(y) {c(y,y[4]*y[5],1);};   # y -> h(y)=(y1,y2,y3,y4,y5,y4*y5,1)
link.temp="logit"
design_p=c(0.0848, 0.0875, 0.0410, 0.0856, 0.0690, 0.0515,
          0.0901, 0.0845, 0.0743, 0.0356, 0.0621, 0.0443,
          0.0090, 0.0794, 0.0157, 0.0380, 0.0455, 0.0022)
det.design=4.552715e-06
paras_lowerbound<-c(0.25,1,-0.3,-0.3,0.1,0.35,-8.0)
paras_upperbound<-c(0.45,2,-0.1,0.0,0.4,0.45,-7.0)
 gjoint_b<- function(x) {
 Func_b<-1/(prod(paras_upperbound-paras_lowerbound))
 ##the prior distributions are follow uniform distribution
return(Func_b)
}
 GLM_Exact_Design(k.continuous=k.continuous,design_x=design_x,
 design_p=design_p,det.design=det.design,p=7,ForLion=FALSE,Integral_based=TRUE,
 joint_Func_b=gjoint_b,Lowerbounds=paras_lowerbound, Upperbounds=paras_upperbound,
 rel.diff=0,L=1,N=100,hfunc=hfunc.temp,link=link.temp)

ForLion documentation built on June 10, 2025, 5:13 p.m.