Nothing
R/continuous_clean.R
In ToxicR: Analyzing Toxicology Dose-Response Data
# #############################################################
# # Function: clean_continuous_analysis
# # By default a continuous analysis constraints the result
# # to be proportional to background mean. This is 'undone'
# # in this function so all values are on the origional scale.
# # info: the fit information for a continuous model
# # log_normal: TRUE if the fit was a lognormal
# # dmodel: Fit information type... to be fixed.
# #############################################################
# clean_continuous_analysis <-function(info,log_normal,dmodel){
#
# c("hill","exp-3","exp-5","power")
# if (dmodel == 1) #hill
# {
# if (log_normal){
# DELTA <- diag(c(info$SCALE,info$SCALE,info$dose_scale,1,1));
# info$EST <- DELTA%*%info$EST;
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# if (ncol(info$Y)==1){
# info$F_INFO[1] = info$F_INFO[1] - sum(log(1/info$Y)) + sum(log(info$SCALE/info$Y))
# }else{
#
# }
# }else{
# #rescale the estimates for normality
# if (nrow(info$EST)==5) {#constant variance
# DELTA <- diag(c(info$SCALE,info$SCALE,info$dose_scale,1,1));
# info$EST <- DELTA%*%info$EST; tmp = info$EST[5] + 2*log(info$SCALE)
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# #fix the constant portion of the likelihood
# if (ncol(info$Y)==1){ #individual data
# info$F_INFO[1] = info$F_INFO[1]+log(exp(0.5*tmp*nrow(info$Y)))-log(exp(0.5*info$EST[5]*nrow(info$Y)))
# info$EST[5] = tmp
# }else{ #summarized data
#
# }
# }else{ #non-constant variance
# DELTA <- diag(c(info$SCALE,info$SCALE,info$dose_scale,1,1,1));
# o_est <- info$EST
# info$EST <- DELTA%*%info$EST;
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# #fix the constant portion of the likelihood
# if (ncol(info$Y)==1){ #individual data
# t_est = info$EST
# mean_t = t_est[1] + t_est[2]*(info$X^t_est[4])/(t_est[3]^t_est[4]+(info$X[,1])^t_est[4])
# mean_o = o_est[1] + o_est[2]*(info$X/max(info$X[,1]))^o_est[4]/(o_est[3]^o_est[4]+(info$X/max(info$X))^o_est[4])
# var_t = exp(info$EST[6]+(2-info$EST[5])*log(info$SCALE))*mean_t^info$EST[5]
# var_o = exp(info$EST[6])*mean_o^info$EST[5]
# info$F_INFO[1] = info$F_INFO[1]+sum(0.5*log(var_t))-sum(0.5*log(var_o))
# info$EST[6] = info$EST[6]+(2-info$EST[5])*log(info$SCALE)
# }else{ #summarized data
#
# }
# }
# }
#
# }
# if (dmodel == 2) #exp-3
# { #lognormality is the easiest
# if (log_normal){
# DELTA <- diag(c(info$SCALE,(1/info$dose_scale),1,1));
# info$EST <- DELTA%*%info$EST;
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# if (ncol(info$Y)==1){ #individual data
# info$F_INFO[1] = info$F_INFO[1] - sum(log(1/info$Y)) + sum(log(info$SCALE/info$Y))
# }else{
#
# }
# }else{
# #rescale the estimates for normality
# if (nrow(info$EST)==4) {#constant variance
# DELTA <- diag(c(info$SCALE,(1/info$dose_scale),1,1));
# info$EST <- DELTA%*%info$EST; tmp = info$EST[4] + 2*log(info$SCALE)
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# #fix the constant portion of the likelihood
# if (ncol(info$Y)==1){ #individual data
# info$F_INFO[1] = info$F_INFO[1]+log(exp(0.5*tmp*nrow(info$Y)))-log(exp(0.5*info$EST[4]*nrow(info$Y)))
# info$EST[4] = tmp
# }else{ #summarized data
#
# }
# }else{ #non-constant variance
# DELTA <- diag(c(info$SCALE,(1/info$dose_scale),1,1,1));
# o_est <- info$EST
# info$EST <- DELTA%*%info$EST;
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# #fix the constant portion of the likelihood
# if (ncol(info$Y)==1){ #individual data
# t_est = info$EST
# mean_t = t_est[1] * exp((t_est[2]*info$Y)^t_est[3])
# mean_o = o_est[1] * exp((o_est[2]*info$Y/max(info$Y))^o_est[3])
# var_t = exp(info$EST[5]+(2-info$EST[4])*log(info$SCALE))*mean_t^info$EST[4]
# var_o = exp(info$EST[5])*mean_o^info$EST[4]
# info$F_INFO[1] = info$F_INFO[1]+sum(0.5*log(var_t))-sum(0.5*log(var_o))
# info$EST[5] = info$EST[5]+(2-info$EST[4])*log(info$SCALE)
# }else{ #summarized data
#
# }
# }
# }
#
# }
# if (dmodel == 3) #exp-5
# {
# if (log_normal){
# DELTA <- diag(c(info$SCALE,(1/info$dose_scale),1,1,1));
# info$EST <- DELTA%*%info$EST;
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# if (ncol(info$Y)==1){ #individual data
# info$F_INFO[1] = info$F_INFO[1] - sum(log(1/info$Y)) + sum(log(info$SCALE/info$Y))
# }else{
#
# }
# }
# else{
# #rescale the estimates for normality
# if (nrow(EST)==5) {#constant variance
# DELTA <- diag(c(info$SCALE,(1/info$dose_scale),1,1,1));
# info$EST <- DELTA%*%info$EST; tmp = info$EST[5] + 2*log(info$SCALE)
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# #fix the constant portion of the likelihood
# if (ncol(info$Y)==1){ #individual data
# info$F_INFO[1] = info$F_INFO[1]+log(exp(0.5*tmp*nrow(info$Y)))-log(exp(0.5*info$EST[5]*nrow(info$Y)))
# info$EST[5] = tmp
# }else{ #summarized data
#
# }
# }else{ #non-constant variance
# DELTA <- diag(c(info$SCALE,(1/info$dose_scale),1,1,1,1));
# o_est <- info$EST
# info$EST <- DELTA%*%info$EST;
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# #fix the constant portion of the likelihood
# if (ncol(info$Y)==1){ #individual data
# t_est = info$EST
# mean_t = t_est[1] *(exp(t_est[3])-(exp(t_est[3])-1)*exp(-(t_est[2]*info$Y)^t_est[4]))
# mean_o = o_est[1] *(exp(o_est[3])-(exp(o_est[3])-1)*exp(-(o_est[2]*info$Y/max(info$Y))^o_est[4]))
# var_t = exp(info$EST[6]+(2-info$EST[5])*log(info$SCALE))*mean_t^info$EST[5]
# var_o = exp(info$EST[6])*mean_o^info$EST[5]
# info$F_INFO[1] = info$F_INFO[1]+sum(0.5*log(var_t))-sum(0.5*log(var_o))
# info$EST[6] = info$EST[6]+(2-info$EST[5])*log(info$SCALE)
# }else{ #summarized data
#
# }
# }
# }
#
# }
# if (dmodel == 4) #power
# {
# if (log_normal){
# DELTA <- diag(c(info$SCALE,info$SCALE*(1/info$dose_scale)^(info$EST[3]),1,1));
# info$EST <- DELTA%*%info$EST;
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# if (ncol(info$Y)==1){ #individual data
# info$F_INFO[1] = info$F_INFO[1] - sum(log(1/info$Y)) + sum(log(info$SCALE/info$Y))
# }else{
#
# }
# }else{
# #rescale the estimates for normality
# if (nrow(info$EST)==4) {#constant variance
# DELTA <- diag(c(info$SCALE,info$SCALE*(1/info$dose_scale)^(info$EST[3]),1,1));
# info$EST <- DELTA%*%info$EST; tmp = info$EST[4] + 2*log(info$SCALE)
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# #fix the constant portion of the likelihood
# if (ncol(info$data)){ #individual data
# info$F_INFO[1] = info$F_INFO[1]+log(exp(0.5*tmp*nrow(info$Y)))-log(exp(0.5*info$EST[4]*nrow(info$Y)))
# info$EST[4] = tmp
# }else{ #summarized data
#
# }
# }else{ #non-constant variance
# DELTA <- diag(c(info$SCALE,info$SCALE*(1/info$dose_scale)^(info$EST[3]),1,1,1));
# o_est <- info$EST
# info$EST <- DELTA%*%info$EST;
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# #fix the constant portion of the likelihood
# if (ncol(info$Y)==1){ #individual data
# t_est = info$EST
# mean_t = t_est[1] + t_est[2]*info$Y^t_est[3]
# mean_o = o_est[1] + o_est[2]*(info$Y/max(info$Y))^o_est[3]
# var_t = exp(info$EST[5]+(2-info$EST[4])*log(info$SCALE))*mean_t^info$EST[4]
# var_o = exp(info$EST[5])*mean_o^info$EST[4]
# info$F_INFO[1] = info$F_INFO[1]+sum(0.5*log(var_t))-sum(0.5*log(var_o))
# info$EST[5] = info$EST[5]+(2-info$EST[4])*log(info$SCALE)
# }else{ #summarized data
#
# }
# }
# }
# }
#
# return(info)
# }
#
# #############################################################
# # Function: clean_continuous_analysis
# # By default a continuous analysis constraints the result
# # to be proportional to background mean. This is 'undone'
# # in this function so all values are on the origional scale.
# # info: the fit information for a continuous model
# # log_normal: TRUE if the fit was a lognormal
# # dmodel: Fit information type... to be fixed.
# #############################################################
# clean_parameters <-function(model,A,B,SCALE,dose_scale,deg,log_normal){
# EST <- A
# COV <- B
#
# dmodel = which(model==c("hill","exp-3","exp-5","power","poly"))
#
# if (dmodel == 1) #hill
# { #rescale the estimates for normality
# if (nrow(EST)==5) {#constant variance
# DELTA <- diag(c(SCALE,SCALE,dose_scale,1,1));
# EST <- DELTA%*%EST;
# if (!log_normal){
# EST[5] = EST[5] + 2*log(SCALE)
# }
# COV <- DELTA%*%COV%*%t(DELTA)
# }else{ #non-constant variance
# DELTA <- diag(c(SCALE,SCALE,dose_scale,1,1,1));
# EST <- DELTA%*%EST;
# COV <- DELTA%*%COV%*%t(DELTA)
# }
# }
# if (dmodel == 2) #exp-3
# { #lognormality is the easiest
# if (nrow(EST)==4) {#constant variance
# DELTA <- diag(c(SCALE,(1/dose_scale),1,1));
# EST <- DELTA%*%EST;
# if (!log_normal){
# EST[4] = EST[4] + 2*log(SCALE)
# }
# COV <- DELTA%*%COV%*%t(DELTA)
# }else{ #non-constant variance
# DELTA <- diag(c(SCALE,(1/dose_scale),1,1,1));
# EST <- DELTA%*%EST;
# COV <- DELTA%*%COV%*%t(DELTA)
# }
# }
# if (dmodel == 3) #exp-5
# {
# if (nrow(EST)==5) {#constant variance
# DELTA <- diag(c(SCALE,(1/dose_scale),1,1,1));
# EST <- DELTA%*%EST;
# if (!log_normal){
# EST[5] = EST[5] + 2*log(SCALE)
# }
# COV <- DELTA%*%COV%*%t(DELTA)
# }else{ #non-constant variance
# DELTA <- diag(c(SCALE,(1/dose_scale),1,1,1,1));
# EST <- DELTA%*%EST;
# COV <- DELTA%*%COV%*%t(DELTA)
# }
# }
# if (dmodel == 4) #power
# {
# if (nrow(EST) == 4) {#constant variance
# DELTA <- diag(c(SCALE,SCALE*(1/dose_scale)^(EST[3]),1,1));
# EST <- DELTA%*%EST;
# if (!log_normal){
# EST[4] = EST[4] + 2*log(SCALE)
# }
# DELTA[3,4] = log(1/dose_scale)*SCALE*(1/dose_scale)^EST[3] #delta method
# COV <- DELTA%*%COV%*%t(DELTA)
# } else{ #non-constant variance
# DELTA <- diag(c(SCALE,SCALE*(1/dose_scale)^(EST[3]),1,1,1));
# EST <- DELTA%*%EST;
# DELTA[3,4] = log(1/dose_scale)*SCALE*(1/dose_scale)^EST[3] #delta method
# COV <- DELTA%*%COV%*%t(DELTA)
# }
# }
# if (dmodel == 5) #polynomial
# {
# if (nrow(EST)==deg+2) {#constant variance
# DELTA <- diag(c(SCALE,SCALE*(1/dose_scale)^(1:deg),1));
# EST <- DELTA%*%EST;
# if (!log_normal){
# EST[4] = EST[4] + 2*log(SCALE)
# }
# COV <- DELTA%*%COV%*%t(DELTA)
# } else{ #non-constant variance
# DELTA <- diag(c(SCALE,SCALE*(1/dose_scale)^(1:deg),1,1));
# EST <- DELTA%*%EST;
# COV <- DELTA%*%COV%*%t(DELTA)
# }
# }
# rval <- list(EST,COV)
# names(rval) <- c("EST","COV")
# return(rval)
# }
Try the ToxicR package in your browser
Any scripts or data that you put into this service are public.
ToxicR documentation built on Dec. 28, 2022, 3:07 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
# #############################################################
# # Function: clean_continuous_analysis
# # By default a continuous analysis constraints the result
# # to be proportional to background mean. This is 'undone'
# # in this function so all values are on the origional scale.
# # info: the fit information for a continuous model
# # log_normal: TRUE if the fit was a lognormal
# # dmodel: Fit information type... to be fixed.
# #############################################################
# clean_continuous_analysis <-function(info,log_normal,dmodel){
#
# c("hill","exp-3","exp-5","power")
# if (dmodel == 1) #hill
# {
# if (log_normal){
# DELTA <- diag(c(info$SCALE,info$SCALE,info$dose_scale,1,1));
# info$EST <- DELTA%*%info$EST;
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# if (ncol(info$Y)==1){
# info$F_INFO[1] = info$F_INFO[1] - sum(log(1/info$Y)) + sum(log(info$SCALE/info$Y))
# }else{
#
# }
# }else{
# #rescale the estimates for normality
# if (nrow(info$EST)==5) {#constant variance
# DELTA <- diag(c(info$SCALE,info$SCALE,info$dose_scale,1,1));
# info$EST <- DELTA%*%info$EST; tmp = info$EST[5] + 2*log(info$SCALE)
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# #fix the constant portion of the likelihood
# if (ncol(info$Y)==1){ #individual data
# info$F_INFO[1] = info$F_INFO[1]+log(exp(0.5*tmp*nrow(info$Y)))-log(exp(0.5*info$EST[5]*nrow(info$Y)))
# info$EST[5] = tmp
# }else{ #summarized data
#
# }
# }else{ #non-constant variance
# DELTA <- diag(c(info$SCALE,info$SCALE,info$dose_scale,1,1,1));
# o_est <- info$EST
# info$EST <- DELTA%*%info$EST;
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# #fix the constant portion of the likelihood
# if (ncol(info$Y)==1){ #individual data
# t_est = info$EST
# mean_t = t_est[1] + t_est[2]*(info$X^t_est[4])/(t_est[3]^t_est[4]+(info$X[,1])^t_est[4])
# mean_o = o_est[1] + o_est[2]*(info$X/max(info$X[,1]))^o_est[4]/(o_est[3]^o_est[4]+(info$X/max(info$X))^o_est[4])
# var_t = exp(info$EST[6]+(2-info$EST[5])*log(info$SCALE))*mean_t^info$EST[5]
# var_o = exp(info$EST[6])*mean_o^info$EST[5]
# info$F_INFO[1] = info$F_INFO[1]+sum(0.5*log(var_t))-sum(0.5*log(var_o))
# info$EST[6] = info$EST[6]+(2-info$EST[5])*log(info$SCALE)
# }else{ #summarized data
#
# }
# }
# }
#
# }
# if (dmodel == 2) #exp-3
# { #lognormality is the easiest
# if (log_normal){
# DELTA <- diag(c(info$SCALE,(1/info$dose_scale),1,1));
# info$EST <- DELTA%*%info$EST;
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# if (ncol(info$Y)==1){ #individual data
# info$F_INFO[1] = info$F_INFO[1] - sum(log(1/info$Y)) + sum(log(info$SCALE/info$Y))
# }else{
#
# }
# }else{
# #rescale the estimates for normality
# if (nrow(info$EST)==4) {#constant variance
# DELTA <- diag(c(info$SCALE,(1/info$dose_scale),1,1));
# info$EST <- DELTA%*%info$EST; tmp = info$EST[4] + 2*log(info$SCALE)
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# #fix the constant portion of the likelihood
# if (ncol(info$Y)==1){ #individual data
# info$F_INFO[1] = info$F_INFO[1]+log(exp(0.5*tmp*nrow(info$Y)))-log(exp(0.5*info$EST[4]*nrow(info$Y)))
# info$EST[4] = tmp
# }else{ #summarized data
#
# }
# }else{ #non-constant variance
# DELTA <- diag(c(info$SCALE,(1/info$dose_scale),1,1,1));
# o_est <- info$EST
# info$EST <- DELTA%*%info$EST;
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# #fix the constant portion of the likelihood
# if (ncol(info$Y)==1){ #individual data
# t_est = info$EST
# mean_t = t_est[1] * exp((t_est[2]*info$Y)^t_est[3])
# mean_o = o_est[1] * exp((o_est[2]*info$Y/max(info$Y))^o_est[3])
# var_t = exp(info$EST[5]+(2-info$EST[4])*log(info$SCALE))*mean_t^info$EST[4]
# var_o = exp(info$EST[5])*mean_o^info$EST[4]
# info$F_INFO[1] = info$F_INFO[1]+sum(0.5*log(var_t))-sum(0.5*log(var_o))
# info$EST[5] = info$EST[5]+(2-info$EST[4])*log(info$SCALE)
# }else{ #summarized data
#
# }
# }
# }
#
# }
# if (dmodel == 3) #exp-5
# {
# if (log_normal){
# DELTA <- diag(c(info$SCALE,(1/info$dose_scale),1,1,1));
# info$EST <- DELTA%*%info$EST;
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# if (ncol(info$Y)==1){ #individual data
# info$F_INFO[1] = info$F_INFO[1] - sum(log(1/info$Y)) + sum(log(info$SCALE/info$Y))
# }else{
#
# }
# }
# else{
# #rescale the estimates for normality
# if (nrow(EST)==5) {#constant variance
# DELTA <- diag(c(info$SCALE,(1/info$dose_scale),1,1,1));
# info$EST <- DELTA%*%info$EST; tmp = info$EST[5] + 2*log(info$SCALE)
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# #fix the constant portion of the likelihood
# if (ncol(info$Y)==1){ #individual data
# info$F_INFO[1] = info$F_INFO[1]+log(exp(0.5*tmp*nrow(info$Y)))-log(exp(0.5*info$EST[5]*nrow(info$Y)))
# info$EST[5] = tmp
# }else{ #summarized data
#
# }
# }else{ #non-constant variance
# DELTA <- diag(c(info$SCALE,(1/info$dose_scale),1,1,1,1));
# o_est <- info$EST
# info$EST <- DELTA%*%info$EST;
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# #fix the constant portion of the likelihood
# if (ncol(info$Y)==1){ #individual data
# t_est = info$EST
# mean_t = t_est[1] *(exp(t_est[3])-(exp(t_est[3])-1)*exp(-(t_est[2]*info$Y)^t_est[4]))
# mean_o = o_est[1] *(exp(o_est[3])-(exp(o_est[3])-1)*exp(-(o_est[2]*info$Y/max(info$Y))^o_est[4]))
# var_t = exp(info$EST[6]+(2-info$EST[5])*log(info$SCALE))*mean_t^info$EST[5]
# var_o = exp(info$EST[6])*mean_o^info$EST[5]
# info$F_INFO[1] = info$F_INFO[1]+sum(0.5*log(var_t))-sum(0.5*log(var_o))
# info$EST[6] = info$EST[6]+(2-info$EST[5])*log(info$SCALE)
# }else{ #summarized data
#
# }
# }
# }
#
# }
# if (dmodel == 4) #power
# {
# if (log_normal){
# DELTA <- diag(c(info$SCALE,info$SCALE*(1/info$dose_scale)^(info$EST[3]),1,1));
# info$EST <- DELTA%*%info$EST;
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# if (ncol(info$Y)==1){ #individual data
# info$F_INFO[1] = info$F_INFO[1] - sum(log(1/info$Y)) + sum(log(info$SCALE/info$Y))
# }else{
#
# }
# }else{
# #rescale the estimates for normality
# if (nrow(info$EST)==4) {#constant variance
# DELTA <- diag(c(info$SCALE,info$SCALE*(1/info$dose_scale)^(info$EST[3]),1,1));
# info$EST <- DELTA%*%info$EST; tmp = info$EST[4] + 2*log(info$SCALE)
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# #fix the constant portion of the likelihood
# if (ncol(info$data)){ #individual data
# info$F_INFO[1] = info$F_INFO[1]+log(exp(0.5*tmp*nrow(info$Y)))-log(exp(0.5*info$EST[4]*nrow(info$Y)))
# info$EST[4] = tmp
# }else{ #summarized data
#
# }
# }else{ #non-constant variance
# DELTA <- diag(c(info$SCALE,info$SCALE*(1/info$dose_scale)^(info$EST[3]),1,1,1));
# o_est <- info$EST
# info$EST <- DELTA%*%info$EST;
# info$COV <- DELTA%*%info$COV%*%t(DELTA)
# #fix the constant portion of the likelihood
# if (ncol(info$Y)==1){ #individual data
# t_est = info$EST
# mean_t = t_est[1] + t_est[2]*info$Y^t_est[3]
# mean_o = o_est[1] + o_est[2]*(info$Y/max(info$Y))^o_est[3]
# var_t = exp(info$EST[5]+(2-info$EST[4])*log(info$SCALE))*mean_t^info$EST[4]
# var_o = exp(info$EST[5])*mean_o^info$EST[4]
# info$F_INFO[1] = info$F_INFO[1]+sum(0.5*log(var_t))-sum(0.5*log(var_o))
# info$EST[5] = info$EST[5]+(2-info$EST[4])*log(info$SCALE)
# }else{ #summarized data
#
# }
# }
# }
# }
#
# return(info)
# }
#
# #############################################################
# # Function: clean_continuous_analysis
# # By default a continuous analysis constraints the result
# # to be proportional to background mean. This is 'undone'
# # in this function so all values are on the origional scale.
# # info: the fit information for a continuous model
# # log_normal: TRUE if the fit was a lognormal
# # dmodel: Fit information type... to be fixed.
# #############################################################
# clean_parameters <-function(model,A,B,SCALE,dose_scale,deg,log_normal){
# EST <- A
# COV <- B
#
# dmodel = which(model==c("hill","exp-3","exp-5","power","poly"))
#
# if (dmodel == 1) #hill
# { #rescale the estimates for normality
# if (nrow(EST)==5) {#constant variance
# DELTA <- diag(c(SCALE,SCALE,dose_scale,1,1));
# EST <- DELTA%*%EST;
# if (!log_normal){
# EST[5] = EST[5] + 2*log(SCALE)
# }
# COV <- DELTA%*%COV%*%t(DELTA)
# }else{ #non-constant variance
# DELTA <- diag(c(SCALE,SCALE,dose_scale,1,1,1));
# EST <- DELTA%*%EST;
# COV <- DELTA%*%COV%*%t(DELTA)
# }
# }
# if (dmodel == 2) #exp-3
# { #lognormality is the easiest
# if (nrow(EST)==4) {#constant variance
# DELTA <- diag(c(SCALE,(1/dose_scale),1,1));
# EST <- DELTA%*%EST;
# if (!log_normal){
# EST[4] = EST[4] + 2*log(SCALE)
# }
# COV <- DELTA%*%COV%*%t(DELTA)
# }else{ #non-constant variance
# DELTA <- diag(c(SCALE,(1/dose_scale),1,1,1));
# EST <- DELTA%*%EST;
# COV <- DELTA%*%COV%*%t(DELTA)
# }
# }
# if (dmodel == 3) #exp-5
# {
# if (nrow(EST)==5) {#constant variance
# DELTA <- diag(c(SCALE,(1/dose_scale),1,1,1));
# EST <- DELTA%*%EST;
# if (!log_normal){
# EST[5] = EST[5] + 2*log(SCALE)
# }
# COV <- DELTA%*%COV%*%t(DELTA)
# }else{ #non-constant variance
# DELTA <- diag(c(SCALE,(1/dose_scale),1,1,1,1));
# EST <- DELTA%*%EST;
# COV <- DELTA%*%COV%*%t(DELTA)
# }
# }
# if (dmodel == 4) #power
# {
# if (nrow(EST) == 4) {#constant variance
# DELTA <- diag(c(SCALE,SCALE*(1/dose_scale)^(EST[3]),1,1));
# EST <- DELTA%*%EST;
# if (!log_normal){
# EST[4] = EST[4] + 2*log(SCALE)
# }
# DELTA[3,4] = log(1/dose_scale)*SCALE*(1/dose_scale)^EST[3] #delta method
# COV <- DELTA%*%COV%*%t(DELTA)
# } else{ #non-constant variance
# DELTA <- diag(c(SCALE,SCALE*(1/dose_scale)^(EST[3]),1,1,1));
# EST <- DELTA%*%EST;
# DELTA[3,4] = log(1/dose_scale)*SCALE*(1/dose_scale)^EST[3] #delta method
# COV <- DELTA%*%COV%*%t(DELTA)
# }
# }
# if (dmodel == 5) #polynomial
# {
# if (nrow(EST)==deg+2) {#constant variance
# DELTA <- diag(c(SCALE,SCALE*(1/dose_scale)^(1:deg),1));
# EST <- DELTA%*%EST;
# if (!log_normal){
# EST[4] = EST[4] + 2*log(SCALE)
# }
# COV <- DELTA%*%COV%*%t(DELTA)
# } else{ #non-constant variance
# DELTA <- diag(c(SCALE,SCALE*(1/dose_scale)^(1:deg),1,1));
# EST <- DELTA%*%EST;
# COV <- DELTA%*%COV%*%t(DELTA)
# }
# }
# rval <- list(EST,COV)
# names(rval) <- c("EST","COV")
# return(rval)
# }
Try the ToxicR package in your browser
Any scripts or data that you put into this service are public.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.