Simulations/brice-biostat-cluster/report/table1.R
In paulowhite/DelayedGSD: Planning and analysis of group sequential trial with pipeline data
### table1.R ---
##----------------------------------------------------------------------
## Author: Brice Ozenne
## Created: okt 19 2023 (10:24)
## Version:
## Last-Updated: okt 19 2023 (17:16)
## By: Brice Ozenne
## Update #: 36
##----------------------------------------------------------------------
##
### Commentary:
##
### Change Log:
##----------------------------------------------------------------------
##
### Code:
if(Sys.info()["login"] == "bozenne"){
}else if(Sys.info()["login"] == "hpl802"){
setwd("x:/DelayedGSD/")
}
options(width = 120, digits = 4)
library(data.table)
library(ggplot2)
library(xtable)
## * formating functions
##' @description Average and nicely format as percentage (fixed number of digits, e.g. 0.010%)
mean2pc <- function(x, digits = 2){
mean.x <- mean(x, na.rm = TRUE)
out <- paste0(formatC(100*mean.x, format='f', digits = digits), "%")
if(round(100*mean.x,digits)==0){
out[round(100*mean.x,digits)==0] <- paste0("<0.",rep(0,digits-1),"1%")
}
if(abs(mean.x)<1e-12){
out[abs(mean.x)<1e-12] <- "0"
}
if(any(is.na(x))){
out <- paste0(out," (NA: ",dec2pc(mean(is.na(x)), digits = digits),")")
}
return(out)
}
##' @description Average and nicely format as numeric (fixed number of digits, e.g. 0.010%)
mean2num <- function(x, digits = 3){
out <- formatC(mean(x, na.rm = TRUE), format='f', digits = digits)
if(any(is.na(x))){
out <- paste0(out," (NA: ",dec2pc(mean(is.na(x)), digits = digits),")")
}
return(out)
}
##' @description Generate table
createTableResSim <- function(data, xtable, sep = " / "){
## prepare data
## select the last interim
data.interim <- data[type == "interim", .(decision.interim=paste0(decision[which.max(stage)]," (",reason[which.max(stage)],")")),
by = c("scenario","seed","method")]
data.decision <- merge(x = data[decision %in% c("futility","efficacy")],
y = data.interim, by = c("scenario","method","seed"),
all = TRUE)
data.decision[, method.char := factor(method.char, levels = c("method 1","method 1 fixC","method 2","method 2 fixC","method 3"))]
setkeyv(data.decision, "method.char")
ls.sumstat <- list()
ls.sumstat$type1 <- data.decision[hypo=="typeI",.(statistic = "type 1 error",
value = mean2pc(decision=="efficacy")),
by="method.char"]
ls.sumstat$power <- data.decision[hypo=="power",.(statistic = "power",
value = mean2pc(decision=="efficacy")),
by="method.char"]
ls.sumstat$CINA <- data.decision[hypo=="power",.(statistic = "CI [NA,NA]",
value = mean2pc(is.na(lower_MUE) | is.na(upper_MUE))),
by="method.char"]
ls.sumstat$coverage <- data.decision[hypo=="power" & !is.na(lower_MUE) & !is.na(upper_MUE),.(statistic = "coverage",
value = mean2pc(lower_MUE <= truth & truth <= upper_MUE)),
by="method.char"]
ls.sumstat$reversal <- data.decision[hypo=="power",.(statistic = "reversal",
value = paste(mean2pc(decision.interim=="stop (futility)" & decision=="efficacy"),
mean2pc(decision.interim=="stop (efficacy)" & decision=="futility"),
sep="/")),
by="method.char"]
ls.sumstat$abnormal <- data.decision[hypo=="power",.(statistic = "abnormal",
value = paste(mean2pc(decision=="efficacy" & statistic < 1.96),
mean2pc(decision=="futility" & statistic >= ck),
sep="/")),
by="method.char"]
ls.sumstat$LMMEmeanB <- data.decision[!is.na(estimate_MUE),.(statistic = "mean bias LMME",
value = paste(mean2num(.SD[hypo=="typeI",estimate_ML]-.SD[hypo=="typeI",truth]),
mean2num(.SD[hypo=="power",estimate_ML]-.SD[hypo=="power",truth]),
sep=sep)),
by="method.char"]
ls.sumstat$MUEmeanB <- data.decision[!is.na(estimate_MUE),.(statistic = "mean bias MUE",
value = paste(mean2num(.SD[hypo=="typeI",estimate_MUE]-.SD[hypo=="typeI",truth]),
mean2num(.SD[hypo=="power",estimate_MUE]-.SD[hypo=="power",truth]),
sep=sep)),
by="method.char"]
ls.sumstat$LMMEmedianB <- data.decision[!is.na(estimate_MUE),.(statistic = "median bias LMME",
value = paste(mean2pc((.SD[hypo=="typeI",estimate_ML]>.SD[hypo=="typeI",truth])-0.5),
mean2pc((.SD[hypo=="power",estimate_ML]>.SD[hypo=="power",truth])-0.5),
sep=sep)),
by="method.char"]
ls.sumstat$MUEmedianB <- data.decision[!is.na(estimate_MUE),.(statistic = "median bias MUE",
value = paste(mean2pc((.SD[hypo=="typeI",estimate_MUE]>.SD[hypo=="typeI",truth])-0.5),
mean2pc((.SD[hypo=="power",estimate_MUE]>.SD[hypo=="power",truth])-0.5),
sep=sep)),
by="method.char"]
dtL.sumstat <- do.call(rbind,ls.sumstat)
dtL.sumstat[, statistic := factor(statistic, unique(statistic))]
dtW.sumstat <- dcast(dtL.sumstat, formula = statistic~method.char, value.var = "value")
if(xtable){
## % -> \\%
dtW.sumstat[[2]] <- gsub("%","\\%",dtW.sumstat[[2]], fixed = TRUE)
dtW.sumstat[[3]] <- gsub("%","\\%",dtW.sumstat[[3]], fixed = TRUE)
dtW.sumstat[[4]] <- gsub("%","\\%",dtW.sumstat[[4]], fixed = TRUE)
dtW.sumstat[[5]] <- gsub("%","\\%",dtW.sumstat[[5]], fixed = TRUE)
dtW.sumstat[[6]] <- gsub("%","\\%",dtW.sumstat[[6]], fixed = TRUE)
add <- "\\hspace{3mm}"
dtW.sumstat$statistic <- paste0(add,c("Type 1 error",
"Power",
"CI=[NA;NA]",
"Coverage",
"Reversal(F\\(\\rightarrow\\)E/E\\(\\rightarrow\\)F)",
"E\\(\\left(\\tilde{Z}_k<1.96\\right)\\)/F\\(\\left(\\tilde{Z}\\geq c_k\\right)\\)",
"Mean bias (\\(\\mathcal{H}_0\\)/\\(\\mathcal{H}_1\\)): LMME",
"\\hphantom{Mean bias (\\(\\mathcal{H}_0\\)/\\(\\mathcal{H}_1\\))}: MUE",
"Median bias (\\(\\mathcal{H}_0\\)/\\(\\mathcal{H}_1\\)): LMME",
"\\hphantom{Median bias (\\(\\mathcal{H}_0\\)/\\(\\mathcal{H}_1\\))}: MUE"))
xtable.sumstat <- xtable(cbind(dtW.sumstat[,1:3],"x"="",dtW.sumstat[,4:5],"y"="",dtW.sumstat[,6]),
type='latex')
add.to.row <- list(pos = list(4,6), command = c("[2mm]","[2mm]"))
print(xtable.sumstat, include.rownames = FALSE, sanitize.text.function=identity, add.to.row = add.to.row)
}else{
return(dtW.sumstat)
}
}
## * 2 stages
## ** Load data
res2stage <- readRDS(file.path("Results-built","res2stage.rds"))
res2stage[, method.char := paste0("method ",method, c(""," fixC")[fixC+1])]
res2stage[, stage.char := factor(stage, 1:2, c("interim","final"))]
res2stage[, truth := ifelse(hypo=="power",0.6,0)]
## ** Generate table
keep.col <- c("scenario", "hypo", "method", "stage", "type", "statistic", "ck",
"estimate_ML", "se_ML", "p.value_ML", "lower_ML", "upper_ML",
"estimate_MUE", "p.value_MUE", "lower_MUE", "upper_MUE",
"decision", "reason", "method.char", "stage.char", "truth","seed")
## *** ar 5 binding
res2stage.ar5binding <- res2stage[method.char != "method 3 fixC" & missing==TRUE & binding==TRUE & ar==5 & !is.na(decision),.SD,.SDcols=keep.col]
res2stage.ar5binding[,.N,by=c("method.char","type","hypo")][type=="interim",unique(N)]
## [1] 10000
table2stage.ar5binding <- createTableResSim(res2stage.ar5binding, xtable = FALSE)
table2stage.ar5binding
## statistic method 1 method 1 fixC method 2 method 2 fixC method 3
## 1: type 1 error 2.40% 2.32% 2.40% 2.31% 2.35%
## 2: power 80.53% 80.08% 80.53% 80.20% 80.14%
## 3: CI [NA,NA] 0.01% 0.01% 0.01% 0.01% 0.01%
## 4: coverage 94.74% 96.29% 94.74% 96.33% 95.14%
## 5: reversal 0.08%/0.07% 0.02%/0.46% 0.08%/0.07% 0.02%/0.45% 0/0.67%
## 6: abnormal 0.45%/0 0/0 0.45%/0 0/0 0/0.02%
## 7: mean bias LMME -0.030/0.023 -0.030/0.023 -0.030/0.023 -0.031/0.023 -0.031/0.024
## 8: mean bias MUE -0.012/0.010 -0.012/-0.015 -0.012/0.010 -0.013/-0.015 -0.012/0.004
## 9: median bias LMME -3.29%/4.05% -3.29%/4.05% -3.28%/4.05% -3.45%/4.07% -3.44%/4.32%
## 10: median bias MUE 0.06%/-0.35% 0.06%/-0.87% 0.06%/-0.34% 0.08%/-0.76% 0.07%/-0.55%
createTableResSim(res2stage.ar5binding, xtable = TRUE)
power2stage.ar5binding <- as.numeric(gsub("%","",unlist(table2stage.ar5binding[2,.SD,.SDcols = names(table2stage.ar5binding)[-1]]),fixed=TRUE))
power2stage.ar5binding[1:2]-power2stage.ar5binding[3:4]
power2stage.ar5binding[1:2]-power2stage.ar5binding[5]
## > [1] 0.00 -0.12
## > [1] 0.39 -0.06
## *** ar 10 binding
res2stage.ar10binding <- res2stage[method.char != "method 3 fixC" & missing==TRUE & binding==TRUE & ar==10 & !is.na(decision),.SD,.SDcols=keep.col]
res2stage.ar10binding[,.N,by=c("method.char","type","hypo")][type=="interim",unique(N)]
## [1] 10000
table2stage.ar10binding <- createTableResSim(res2stage.ar10binding, xtable = FALSE)
table2stage.ar10binding
## statistic method 1 method 1 fixC method 2 method 2 fixC method 3
## 1: type 1 error 2.42% 2.24% 2.39% 2.22% 2.37%
## 2: power 81.00% 80.15% 80.93% 80.35% 80.43%
## 3: CI [NA,NA] 0.02% 0.02% 0.02% 0.02% 0.02%
## 4: coverage 94.84% 96.26% 94.82% 96.31% 95.34%
## 5: reversal 0.57%/0.17% 0.22%/0.67% 0.61%/0.20% 0.16%/0.65% 0/1.07%
## 6: abnormal 0.85%/0 0/0 0.84%/0 0/0 0/0.11%
## 7: mean bias LMME -0.018/0.013 -0.018/0.013 -0.018/0.013 -0.019/0.014 -0.019/0.015
## 8: mean bias MUE -0.005/0.006 -0.006/-0.015 -0.004/0.006 -0.006/-0.015 -0.005/-0.003
## 9: median bias LMME -1.72%/2.61% -1.72%/2.61% -1.69%/2.60% -1.97%/2.65% -2.02%/3.01%
## 10: median bias MUE 0.10%/-0.24% 0.10%/-1.11% 0.08%/-0.25% -0.07%/-1.05% -0.02%/-0.55%
createTableResSim(res2stage.ar10binding, xtable = TRUE)
power2stage.ar10binding <- as.numeric(gsub("%","",unlist(table2stage.ar10binding[2,.SD,.SDcols = names(table2stage.ar10binding)[-1]]),fixed=TRUE))
power2stage.ar10binding[1:2]-power2stage.ar10binding[3:4]
power2stage.ar10binding[1:2]-power2stage.ar10binding[5]
## [1] 0.07 -0.20
## [1] 0.57 -0.28
## *** ar 5 non-binding
res2stage.ar5nonbinding <- res2stage[method.char != "method 3 fixC" & missing==TRUE & binding==FALSE & ar==5 & !is.na(decision),.SD,.SDcols=keep.col]
res2stage.ar5nonbinding[,.N,by=c("method.char","type","hypo")][type=="interim",unique(N)]
## [1] 10000
table2stage.ar5nonbinding <- createTableResSim(res2stage.ar5nonbinding, xtable = FALSE)
table2stage.ar5nonbinding
## statistic method 1 method 1 fixC method 2 method 2 fixC method 3
## 1: type 1 error 2.68% 2.63% 2.68% 2.64% 2.66%
## 2: power 80.37% 79.93% 80.36% 80.04% 80.06%
## 3: CI [NA,NA] 5.72% 6.16% 5.76% 5.86% 6.05%
## 4: coverage 95.86% 97.77% 95.86% 97.76% 96.55%
## 5: reversal 0.03%/0.04% 0.01%/0.46% 0.03%/0.04% 0.01%/0.44% 0/0.60%
## 6: abnormal 0.44%/0 0/0 0.44%/0 0/0 0/0
## 7: mean bias LMME 0.001/0.054 0.001/0.055 0.001/0.054 0.001/0.054 0.001/0.055
## 8: mean bias MUE 0.001/0.042 0.000/0.017 0.001/0.042 0.000/0.015 0.001/0.037
## 9: median bias LMME -0.10%/7.92% -0.12%/8.19% -0.10%/7.97% -0.13%/7.97% -0.14%/8.40%
## 10: median bias MUE -0.11%/2.83% -0.13%/2.57% -0.11%/2.87% -0.14%/2.47% -0.02%/2.80%
createTableResSim(res2stage.ar5nonbinding, xtable = TRUE)
power2stage.ar5nonbinding <- as.numeric(gsub("%","",unlist(table2stage.ar5nonbinding[2,.SD,.SDcols = names(table2stage.ar5nonbinding)[-1]]),fixed=TRUE))
power2stage.ar5nonbinding[1:2]-power2stage.ar5nonbinding[3:4]
power2stage.ar5nonbinding[1:2]-power2stage.ar5nonbinding[5]
## [1] 0.01 -0.11
## [1] 0.31 -0.13
## *** ar 10 non-binding
res2stage.ar10nonbinding <- res2stage[method.char != "method 3 fixC" & missing==TRUE & binding==FALSE & ar==10 & !is.na(decision),.SD,.SDcols=keep.col]
res2stage.ar10nonbinding[,.N,by=c("method.char","type","hypo")][type=="interim",unique(N)]
## [1] 10000
table2stage.ar10nonbinding <- createTableResSim(res2stage.ar10nonbinding, xtable = FALSE)
table2stage.ar10nonbinding
## statistic method 1 method 1 fixC method 2 method 2 fixC method 3
## 1: type 1 error 2.53% 2.45% 2.53% 2.47% 2.57%
## 2: power 80.50% 79.86% 80.44% 80.12% 80.26%
## 3: CI [NA,NA] 5.95% 6.59% 6.11% 6.06% 6.21%
## 4: coverage 95.90% 97.38% 95.89% 97.45% 97.00%
## 5: reversal 0.41%/0.21% 0.14%/0.58% 0.42%/0.22% 0.11%/0.55% 0/1.04%
## 6: abnormal 0.64%/0 0/0 0.64%/0 0/0 0/0.08%
## 7: mean bias LMME -0.000/0.042 -0.001/0.043 -0.000/0.042 -0.001/0.042 -0.001/0.042
## 8: mean bias MUE -0.000/0.036 -0.001/0.017 -0.000/0.036 -0.001/0.015 0.001/0.029
## 9: median bias LMME -0.15%/6.61% -0.19%/7.00% -0.15%/6.69% -0.19%/6.74% -0.26%/6.79%
## 10: median bias MUE -0.15%/3.09% -0.19%/2.72% -0.14%/3.18% -0.19%/2.47% -0.03%/2.80%
createTableResSim(res2stage.ar10nonbinding, xtable = TRUE)
power2stage.ar10nonbinding <- as.numeric(gsub("%","",unlist(table2stage.ar10nonbinding[2,.SD,.SDcols = names(table2stage.ar10nonbinding)[-1]]),fixed=TRUE))
power2stage.ar10nonbinding[1:2]-power2stage.ar10nonbinding[3:4]
power2stage.ar10nonbinding[1:2]-power2stage.ar10nonbinding[5]
## [1] 0.06 -0.26
## [1] 0.24 -0.40
## * 3 stages
## ** Load data
res3stage <- readRDS(file.path("Results-built","res3stage.rds"))
res3stage[, method.char := paste0("method ",method, c(""," fixC")[fixC+1])]
res3stage[, stage.char := factor(stage, 1:3, c("interim1","interim2","final"))]
res3stage[, truth := ifelse(hypo=="power",0.6,0)]
## ** Generate table
keep.col <- c("scenario", "hypo", "method", "stage", "type", "statistic", "ck",
"estimate_ML", "se_ML", "p.value_ML", "lower_ML", "upper_ML",
"estimate_MUE", "p.value_MUE", "lower_MUE", "upper_MUE",
"decision", "reason", "method.char", "stage.char", "truth","seed")
## *** ar 5 binding
res3stage.ar5binding <- res3stage[method.char != "method 3 fixC" & missing==TRUE & binding==TRUE & ar==5 & !is.na(decision),.SD,.SDcols=keep.col]
res3stage.ar5binding[,.N,by=c("method.char","type","hypo")][type=="interim",unique(N)]
## [1] 10000
table3stage.ar5binding <- createTableResSim(res3stage.ar5binding, xtable = FALSE)
table3stage.ar5binding
## statistic method 1 method 1 fixC method 2 method 2 fixC method 3
## 1: type 1 error 2.61% 2.56% 2.61% 2.55% 2.59%
## 2: power 74.35% 74.00% 74.35% 74.01% 73.99%
## 3: CI [NA,NA] 0 0.08% 0 0.08% 0.11%
## 4: coverage 94.63% 96.84% 94.63% 96.82% 95.68%
## 5: reversal 0.01%/0.02% 0/0.36% 0.01%/0.02% 0/0.35% 0/0.45%
## 6: abnormal 0.35%/0 0/0 0.35%/0 0/0 0/0
## 7: mean bias LMME -0.056 / 0.034 -0.056 / 0.034 -0.056 / 0.034 -0.056 / 0.035 -0.056 / 0.035
## 8: mean bias MUE -0.034 / 0.026 -0.034 / 0.005 -0.034 / 0.026 -0.034 / 0.006 -0.034 / 0.016
## 9: median bias LMME -6.59% / 3.91% -6.60% / 3.89% -6.57% / 3.90% -6.59% / 3.85% -6.60% / 3.97%
## 10: median bias MUE -0.15% / 0.81% -0.16% / 0.50% -0.12% / 0.80% -0.12% / 0.53% -0.13% / 0.53%
createTableResSim(res3stage.ar5binding, xtable = TRUE)
## *** ar 10 binding
res3stage.ar10binding <- res3stage[method.char != "method 3 fixC" & missing==TRUE & binding==TRUE & ar==10 & !is.na(decision),.SD,.SDcols=keep.col]
res3stage.ar10binding[,.N,by=c("method.char","type","hypo")][type=="interim",unique(N)]
## [1] 10000
table3stage.ar10binding <- createTableResSim(res3stage.ar10binding, xtable = FALSE)
table3stage.ar10binding
## statistic method 1 method 1 fixC method 2 method 2 fixC method 3
## 1: type 1 error 2.60% 2.47% 2.60% 2.50% 2.49%
## 2: power 74.51% 73.84% 74.51% 73.91% 74.01%
## 3: CI [NA,NA] 0.12% 0.20% 0.12% 0.20% 0.23%
## 4: coverage 95.07% 96.87% 95.07% 96.91% 96.12%
## 5: reversal 0.25%/0.08% 0.04%/0.54% 0.25%/0.08% 0.04%/0.50% 0/0.73%
## 6: abnormal 0.67%/0 0/0 0.67%/0 0/0 0/0.03%
## 7: mean bias LMME -0.035 / 0.021 -0.035 / 0.021 -0.035 / 0.021 -0.034 / 0.022 -0.034 / 0.023
## 8: mean bias MUE -0.027 / 0.023 -0.028 / 0.008 -0.027 / 0.023 -0.028 / 0.008 -0.028 / 0.014
## 9: median bias LMME -3.89% / 2.84% -3.91% / 2.85% -3.88% / 2.84% -3.80% / 2.85% -3.80% / 3.07%
## 10: median bias MUE -0.01% / 0.72% -0.05% / 0.13% 0.02% / 0.70% 0.08% / 0.19% 0.13% / 0.25%
createTableResSim(res3stage.ar10binding, xtable = TRUE)
## *** ar 5 non-binding
res3stage.ar5nonbinding <- res3stage[method.char != "method 3 fixC" & missing==TRUE & binding==FALSE & ar==5 & !is.na(decision),.SD,.SDcols=keep.col]
res3stage.ar5nonbinding[,.N,by=c("method.char","type","hypo")][type=="interim",unique(N)]
## [1] 10000
table3stage.ar5nonbinding <- createTableResSim(res3stage.ar5nonbinding, xtable = FALSE)
table3stage.ar5nonbinding
## statistic method 1 method 1 fixC method 2 method 2 fixC method 3
## 1: type 1 error 2.53% 2.43% 2.53% 2.44% 2.44%
## 2: power 74.61% 74.33% 74.59% 74.37% 74.43%
## 3: CI [NA,NA] 4.85% 5.13% 4.86% 5.00% 5.09%
## 4: coverage 95.87% 98.24% 95.89% 98.24% 97.00%
## 5: reversal 0.01%/0.01% 0.01%/0.29% 0.01%/0.01% 0.01%/0.28% 0/0.32%
## 6: abnormal 0.28%/0 0/0 0.28%/0 0/0 0/0.01%
## 7: mean bias LMME 0.004 / 0.065 0.003 / 0.065 0.004 / 0.065 0.003 / 0.065 0.003 / 0.066
## 8: mean bias MUE 0.004 / 0.057 0.003 / 0.038 0.004 / 0.057 0.003 / 0.037 0.004 / 0.049
## 9: median bias LMME -0.35% / 5.44% -0.40% / 5.54% -0.35% / 5.44% -0.40% / 5.43% -0.41% / 5.62%
## 10: median bias MUE -0.34% / 2.55% -0.39% / 2.41% -0.35% / 2.55% -0.39% / 2.35% -0.39% / 2.52%
createTableResSim(res3stage.ar5nonbinding, xtable = TRUE)
## *** ar 10 non-binding
res3stage.ar10nonbinding <- res3stage[method.char != "method 3 fixC" & missing==TRUE & binding==FALSE & ar==10 & !is.na(decision),.SD,.SDcols=keep.col]
res3stage.ar10nonbinding[,.N,by=c("method.char","type","hypo")][type=="interim",unique(N)]
## [1] 10000
table3stage.ar10nonbinding <- createTableResSim(res3stage.ar10nonbinding, xtable = FALSE)
table3stage.ar10nonbinding
## statistic method 1 method 1 fixC method 2 method 2 fixC method 3
## 1: type 1 error 2.49% 2.37% 2.49% 2.37% 2.42%
## 2: power 74.71% 74.17% 74.71% 74.25% 74.45%
## 3: CI [NA,NA] 4.78% 5.25% 4.79% 4.88% 4.99%
## 4: coverage 96.02% 97.96% 96.01% 97.95% 97.41%
## 5: reversal 0.21%/0.09% 0.05%/0.47% 0.21%/0.09% 0.04%/0.46% 0/0.57%
## 6: abnormal 0.54%/0 0/0 0.54%/0 0/0 0/0.03%
## 7: mean bias LMME 0.002 / 0.046 0.002 / 0.046 0.002 / 0.046 0.002 / 0.045 0.002 / 0.046
## 8: mean bias MUE 0.003 / 0.051 0.002 / 0.037 0.003 / 0.051 0.002 / 0.036 0.003 / 0.042
## 9: median bias LMME -0.40% / 4.29% -0.45% / 4.60% -0.40% / 4.29% -0.45% / 4.37% -0.47% / 4.58%
## 10: median bias MUE -0.40% / 2.60% -0.45% / 2.33% -0.40% / 2.59% -0.45% / 2.09% -0.39% / 2.21%
createTableResSim(res3stage.ar10nonbinding, xtable = TRUE)
##----------------------------------------------------------------------
### table1.R ends here
paulowhite/DelayedGSD documentation built on Nov. 1, 2023, 5:36 p.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.
### table1.R ---
##----------------------------------------------------------------------
## Author: Brice Ozenne
## Created: okt 19 2023 (10:24)
## Version:
## Last-Updated: okt 19 2023 (17:16)
## By: Brice Ozenne
## Update #: 36
##----------------------------------------------------------------------
##
### Commentary:
##
### Change Log:
##----------------------------------------------------------------------
##
### Code:
if(Sys.info()["login"] == "bozenne"){
}else if(Sys.info()["login"] == "hpl802"){
setwd("x:/DelayedGSD/")
}
options(width = 120, digits = 4)
library(data.table)
library(ggplot2)
library(xtable)
## * formating functions
##' @description Average and nicely format as percentage (fixed number of digits, e.g. 0.010%)
mean2pc <- function(x, digits = 2){
mean.x <- mean(x, na.rm = TRUE)
out <- paste0(formatC(100*mean.x, format='f', digits = digits), "%")
if(round(100*mean.x,digits)==0){
out[round(100*mean.x,digits)==0] <- paste0("<0.",rep(0,digits-1),"1%")
}
if(abs(mean.x)<1e-12){
out[abs(mean.x)<1e-12] <- "0"
}
if(any(is.na(x))){
out <- paste0(out," (NA: ",dec2pc(mean(is.na(x)), digits = digits),")")
}
return(out)
}
##' @description Average and nicely format as numeric (fixed number of digits, e.g. 0.010%)
mean2num <- function(x, digits = 3){
out <- formatC(mean(x, na.rm = TRUE), format='f', digits = digits)
if(any(is.na(x))){
out <- paste0(out," (NA: ",dec2pc(mean(is.na(x)), digits = digits),")")
}
return(out)
}
##' @description Generate table
createTableResSim <- function(data, xtable, sep = " / "){
## prepare data
## select the last interim
data.interim <- data[type == "interim", .(decision.interim=paste0(decision[which.max(stage)]," (",reason[which.max(stage)],")")),
by = c("scenario","seed","method")]
data.decision <- merge(x = data[decision %in% c("futility","efficacy")],
y = data.interim, by = c("scenario","method","seed"),
all = TRUE)
data.decision[, method.char := factor(method.char, levels = c("method 1","method 1 fixC","method 2","method 2 fixC","method 3"))]
setkeyv(data.decision, "method.char")
ls.sumstat <- list()
ls.sumstat$type1 <- data.decision[hypo=="typeI",.(statistic = "type 1 error",
value = mean2pc(decision=="efficacy")),
by="method.char"]
ls.sumstat$power <- data.decision[hypo=="power",.(statistic = "power",
value = mean2pc(decision=="efficacy")),
by="method.char"]
ls.sumstat$CINA <- data.decision[hypo=="power",.(statistic = "CI [NA,NA]",
value = mean2pc(is.na(lower_MUE) | is.na(upper_MUE))),
by="method.char"]
ls.sumstat$coverage <- data.decision[hypo=="power" & !is.na(lower_MUE) & !is.na(upper_MUE),.(statistic = "coverage",
value = mean2pc(lower_MUE <= truth & truth <= upper_MUE)),
by="method.char"]
ls.sumstat$reversal <- data.decision[hypo=="power",.(statistic = "reversal",
value = paste(mean2pc(decision.interim=="stop (futility)" & decision=="efficacy"),
mean2pc(decision.interim=="stop (efficacy)" & decision=="futility"),
sep="/")),
by="method.char"]
ls.sumstat$abnormal <- data.decision[hypo=="power",.(statistic = "abnormal",
value = paste(mean2pc(decision=="efficacy" & statistic < 1.96),
mean2pc(decision=="futility" & statistic >= ck),
sep="/")),
by="method.char"]
ls.sumstat$LMMEmeanB <- data.decision[!is.na(estimate_MUE),.(statistic = "mean bias LMME",
value = paste(mean2num(.SD[hypo=="typeI",estimate_ML]-.SD[hypo=="typeI",truth]),
mean2num(.SD[hypo=="power",estimate_ML]-.SD[hypo=="power",truth]),
sep=sep)),
by="method.char"]
ls.sumstat$MUEmeanB <- data.decision[!is.na(estimate_MUE),.(statistic = "mean bias MUE",
value = paste(mean2num(.SD[hypo=="typeI",estimate_MUE]-.SD[hypo=="typeI",truth]),
mean2num(.SD[hypo=="power",estimate_MUE]-.SD[hypo=="power",truth]),
sep=sep)),
by="method.char"]
ls.sumstat$LMMEmedianB <- data.decision[!is.na(estimate_MUE),.(statistic = "median bias LMME",
value = paste(mean2pc((.SD[hypo=="typeI",estimate_ML]>.SD[hypo=="typeI",truth])-0.5),
mean2pc((.SD[hypo=="power",estimate_ML]>.SD[hypo=="power",truth])-0.5),
sep=sep)),
by="method.char"]
ls.sumstat$MUEmedianB <- data.decision[!is.na(estimate_MUE),.(statistic = "median bias MUE",
value = paste(mean2pc((.SD[hypo=="typeI",estimate_MUE]>.SD[hypo=="typeI",truth])-0.5),
mean2pc((.SD[hypo=="power",estimate_MUE]>.SD[hypo=="power",truth])-0.5),
sep=sep)),
by="method.char"]
dtL.sumstat <- do.call(rbind,ls.sumstat)
dtL.sumstat[, statistic := factor(statistic, unique(statistic))]
dtW.sumstat <- dcast(dtL.sumstat, formula = statistic~method.char, value.var = "value")
if(xtable){
## % -> \\%
dtW.sumstat[[2]] <- gsub("%","\\%",dtW.sumstat[[2]], fixed = TRUE)
dtW.sumstat[[3]] <- gsub("%","\\%",dtW.sumstat[[3]], fixed = TRUE)
dtW.sumstat[[4]] <- gsub("%","\\%",dtW.sumstat[[4]], fixed = TRUE)
dtW.sumstat[[5]] <- gsub("%","\\%",dtW.sumstat[[5]], fixed = TRUE)
dtW.sumstat[[6]] <- gsub("%","\\%",dtW.sumstat[[6]], fixed = TRUE)
add <- "\\hspace{3mm}"
dtW.sumstat$statistic <- paste0(add,c("Type 1 error",
"Power",
"CI=[NA;NA]",
"Coverage",
"Reversal(F\\(\\rightarrow\\)E/E\\(\\rightarrow\\)F)",
"E\\(\\left(\\tilde{Z}_k<1.96\\right)\\)/F\\(\\left(\\tilde{Z}\\geq c_k\\right)\\)",
"Mean bias (\\(\\mathcal{H}_0\\)/\\(\\mathcal{H}_1\\)): LMME",
"\\hphantom{Mean bias (\\(\\mathcal{H}_0\\)/\\(\\mathcal{H}_1\\))}: MUE",
"Median bias (\\(\\mathcal{H}_0\\)/\\(\\mathcal{H}_1\\)): LMME",
"\\hphantom{Median bias (\\(\\mathcal{H}_0\\)/\\(\\mathcal{H}_1\\))}: MUE"))
xtable.sumstat <- xtable(cbind(dtW.sumstat[,1:3],"x"="",dtW.sumstat[,4:5],"y"="",dtW.sumstat[,6]),
type='latex')
add.to.row <- list(pos = list(4,6), command = c("[2mm]","[2mm]"))
print(xtable.sumstat, include.rownames = FALSE, sanitize.text.function=identity, add.to.row = add.to.row)
}else{
return(dtW.sumstat)
}
}
## * 2 stages
## ** Load data
res2stage <- readRDS(file.path("Results-built","res2stage.rds"))
res2stage[, method.char := paste0("method ",method, c(""," fixC")[fixC+1])]
res2stage[, stage.char := factor(stage, 1:2, c("interim","final"))]
res2stage[, truth := ifelse(hypo=="power",0.6,0)]
## ** Generate table
keep.col <- c("scenario", "hypo", "method", "stage", "type", "statistic", "ck",
"estimate_ML", "se_ML", "p.value_ML", "lower_ML", "upper_ML",
"estimate_MUE", "p.value_MUE", "lower_MUE", "upper_MUE",
"decision", "reason", "method.char", "stage.char", "truth","seed")
## *** ar 5 binding
res2stage.ar5binding <- res2stage[method.char != "method 3 fixC" & missing==TRUE & binding==TRUE & ar==5 & !is.na(decision),.SD,.SDcols=keep.col]
res2stage.ar5binding[,.N,by=c("method.char","type","hypo")][type=="interim",unique(N)]
## [1] 10000
table2stage.ar5binding <- createTableResSim(res2stage.ar5binding, xtable = FALSE)
table2stage.ar5binding
## statistic method 1 method 1 fixC method 2 method 2 fixC method 3
## 1: type 1 error 2.40% 2.32% 2.40% 2.31% 2.35%
## 2: power 80.53% 80.08% 80.53% 80.20% 80.14%
## 3: CI [NA,NA] 0.01% 0.01% 0.01% 0.01% 0.01%
## 4: coverage 94.74% 96.29% 94.74% 96.33% 95.14%
## 5: reversal 0.08%/0.07% 0.02%/0.46% 0.08%/0.07% 0.02%/0.45% 0/0.67%
## 6: abnormal 0.45%/0 0/0 0.45%/0 0/0 0/0.02%
## 7: mean bias LMME -0.030/0.023 -0.030/0.023 -0.030/0.023 -0.031/0.023 -0.031/0.024
## 8: mean bias MUE -0.012/0.010 -0.012/-0.015 -0.012/0.010 -0.013/-0.015 -0.012/0.004
## 9: median bias LMME -3.29%/4.05% -3.29%/4.05% -3.28%/4.05% -3.45%/4.07% -3.44%/4.32%
## 10: median bias MUE 0.06%/-0.35% 0.06%/-0.87% 0.06%/-0.34% 0.08%/-0.76% 0.07%/-0.55%
createTableResSim(res2stage.ar5binding, xtable = TRUE)
power2stage.ar5binding <- as.numeric(gsub("%","",unlist(table2stage.ar5binding[2,.SD,.SDcols = names(table2stage.ar5binding)[-1]]),fixed=TRUE))
power2stage.ar5binding[1:2]-power2stage.ar5binding[3:4]
power2stage.ar5binding[1:2]-power2stage.ar5binding[5]
## > [1] 0.00 -0.12
## > [1] 0.39 -0.06
## *** ar 10 binding
res2stage.ar10binding <- res2stage[method.char != "method 3 fixC" & missing==TRUE & binding==TRUE & ar==10 & !is.na(decision),.SD,.SDcols=keep.col]
res2stage.ar10binding[,.N,by=c("method.char","type","hypo")][type=="interim",unique(N)]
## [1] 10000
table2stage.ar10binding <- createTableResSim(res2stage.ar10binding, xtable = FALSE)
table2stage.ar10binding
## statistic method 1 method 1 fixC method 2 method 2 fixC method 3
## 1: type 1 error 2.42% 2.24% 2.39% 2.22% 2.37%
## 2: power 81.00% 80.15% 80.93% 80.35% 80.43%
## 3: CI [NA,NA] 0.02% 0.02% 0.02% 0.02% 0.02%
## 4: coverage 94.84% 96.26% 94.82% 96.31% 95.34%
## 5: reversal 0.57%/0.17% 0.22%/0.67% 0.61%/0.20% 0.16%/0.65% 0/1.07%
## 6: abnormal 0.85%/0 0/0 0.84%/0 0/0 0/0.11%
## 7: mean bias LMME -0.018/0.013 -0.018/0.013 -0.018/0.013 -0.019/0.014 -0.019/0.015
## 8: mean bias MUE -0.005/0.006 -0.006/-0.015 -0.004/0.006 -0.006/-0.015 -0.005/-0.003
## 9: median bias LMME -1.72%/2.61% -1.72%/2.61% -1.69%/2.60% -1.97%/2.65% -2.02%/3.01%
## 10: median bias MUE 0.10%/-0.24% 0.10%/-1.11% 0.08%/-0.25% -0.07%/-1.05% -0.02%/-0.55%
createTableResSim(res2stage.ar10binding, xtable = TRUE)
power2stage.ar10binding <- as.numeric(gsub("%","",unlist(table2stage.ar10binding[2,.SD,.SDcols = names(table2stage.ar10binding)[-1]]),fixed=TRUE))
power2stage.ar10binding[1:2]-power2stage.ar10binding[3:4]
power2stage.ar10binding[1:2]-power2stage.ar10binding[5]
## [1] 0.07 -0.20
## [1] 0.57 -0.28
## *** ar 5 non-binding
res2stage.ar5nonbinding <- res2stage[method.char != "method 3 fixC" & missing==TRUE & binding==FALSE & ar==5 & !is.na(decision),.SD,.SDcols=keep.col]
res2stage.ar5nonbinding[,.N,by=c("method.char","type","hypo")][type=="interim",unique(N)]
## [1] 10000
table2stage.ar5nonbinding <- createTableResSim(res2stage.ar5nonbinding, xtable = FALSE)
table2stage.ar5nonbinding
## statistic method 1 method 1 fixC method 2 method 2 fixC method 3
## 1: type 1 error 2.68% 2.63% 2.68% 2.64% 2.66%
## 2: power 80.37% 79.93% 80.36% 80.04% 80.06%
## 3: CI [NA,NA] 5.72% 6.16% 5.76% 5.86% 6.05%
## 4: coverage 95.86% 97.77% 95.86% 97.76% 96.55%
## 5: reversal 0.03%/0.04% 0.01%/0.46% 0.03%/0.04% 0.01%/0.44% 0/0.60%
## 6: abnormal 0.44%/0 0/0 0.44%/0 0/0 0/0
## 7: mean bias LMME 0.001/0.054 0.001/0.055 0.001/0.054 0.001/0.054 0.001/0.055
## 8: mean bias MUE 0.001/0.042 0.000/0.017 0.001/0.042 0.000/0.015 0.001/0.037
## 9: median bias LMME -0.10%/7.92% -0.12%/8.19% -0.10%/7.97% -0.13%/7.97% -0.14%/8.40%
## 10: median bias MUE -0.11%/2.83% -0.13%/2.57% -0.11%/2.87% -0.14%/2.47% -0.02%/2.80%
createTableResSim(res2stage.ar5nonbinding, xtable = TRUE)
power2stage.ar5nonbinding <- as.numeric(gsub("%","",unlist(table2stage.ar5nonbinding[2,.SD,.SDcols = names(table2stage.ar5nonbinding)[-1]]),fixed=TRUE))
power2stage.ar5nonbinding[1:2]-power2stage.ar5nonbinding[3:4]
power2stage.ar5nonbinding[1:2]-power2stage.ar5nonbinding[5]
## [1] 0.01 -0.11
## [1] 0.31 -0.13
## *** ar 10 non-binding
res2stage.ar10nonbinding <- res2stage[method.char != "method 3 fixC" & missing==TRUE & binding==FALSE & ar==10 & !is.na(decision),.SD,.SDcols=keep.col]
res2stage.ar10nonbinding[,.N,by=c("method.char","type","hypo")][type=="interim",unique(N)]
## [1] 10000
table2stage.ar10nonbinding <- createTableResSim(res2stage.ar10nonbinding, xtable = FALSE)
table2stage.ar10nonbinding
## statistic method 1 method 1 fixC method 2 method 2 fixC method 3
## 1: type 1 error 2.53% 2.45% 2.53% 2.47% 2.57%
## 2: power 80.50% 79.86% 80.44% 80.12% 80.26%
## 3: CI [NA,NA] 5.95% 6.59% 6.11% 6.06% 6.21%
## 4: coverage 95.90% 97.38% 95.89% 97.45% 97.00%
## 5: reversal 0.41%/0.21% 0.14%/0.58% 0.42%/0.22% 0.11%/0.55% 0/1.04%
## 6: abnormal 0.64%/0 0/0 0.64%/0 0/0 0/0.08%
## 7: mean bias LMME -0.000/0.042 -0.001/0.043 -0.000/0.042 -0.001/0.042 -0.001/0.042
## 8: mean bias MUE -0.000/0.036 -0.001/0.017 -0.000/0.036 -0.001/0.015 0.001/0.029
## 9: median bias LMME -0.15%/6.61% -0.19%/7.00% -0.15%/6.69% -0.19%/6.74% -0.26%/6.79%
## 10: median bias MUE -0.15%/3.09% -0.19%/2.72% -0.14%/3.18% -0.19%/2.47% -0.03%/2.80%
createTableResSim(res2stage.ar10nonbinding, xtable = TRUE)
power2stage.ar10nonbinding <- as.numeric(gsub("%","",unlist(table2stage.ar10nonbinding[2,.SD,.SDcols = names(table2stage.ar10nonbinding)[-1]]),fixed=TRUE))
power2stage.ar10nonbinding[1:2]-power2stage.ar10nonbinding[3:4]
power2stage.ar10nonbinding[1:2]-power2stage.ar10nonbinding[5]
## [1] 0.06 -0.26
## [1] 0.24 -0.40
## * 3 stages
## ** Load data
res3stage <- readRDS(file.path("Results-built","res3stage.rds"))
res3stage[, method.char := paste0("method ",method, c(""," fixC")[fixC+1])]
res3stage[, stage.char := factor(stage, 1:3, c("interim1","interim2","final"))]
res3stage[, truth := ifelse(hypo=="power",0.6,0)]
## ** Generate table
keep.col <- c("scenario", "hypo", "method", "stage", "type", "statistic", "ck",
"estimate_ML", "se_ML", "p.value_ML", "lower_ML", "upper_ML",
"estimate_MUE", "p.value_MUE", "lower_MUE", "upper_MUE",
"decision", "reason", "method.char", "stage.char", "truth","seed")
## *** ar 5 binding
res3stage.ar5binding <- res3stage[method.char != "method 3 fixC" & missing==TRUE & binding==TRUE & ar==5 & !is.na(decision),.SD,.SDcols=keep.col]
res3stage.ar5binding[,.N,by=c("method.char","type","hypo")][type=="interim",unique(N)]
## [1] 10000
table3stage.ar5binding <- createTableResSim(res3stage.ar5binding, xtable = FALSE)
table3stage.ar5binding
## statistic method 1 method 1 fixC method 2 method 2 fixC method 3
## 1: type 1 error 2.61% 2.56% 2.61% 2.55% 2.59%
## 2: power 74.35% 74.00% 74.35% 74.01% 73.99%
## 3: CI [NA,NA] 0 0.08% 0 0.08% 0.11%
## 4: coverage 94.63% 96.84% 94.63% 96.82% 95.68%
## 5: reversal 0.01%/0.02% 0/0.36% 0.01%/0.02% 0/0.35% 0/0.45%
## 6: abnormal 0.35%/0 0/0 0.35%/0 0/0 0/0
## 7: mean bias LMME -0.056 / 0.034 -0.056 / 0.034 -0.056 / 0.034 -0.056 / 0.035 -0.056 / 0.035
## 8: mean bias MUE -0.034 / 0.026 -0.034 / 0.005 -0.034 / 0.026 -0.034 / 0.006 -0.034 / 0.016
## 9: median bias LMME -6.59% / 3.91% -6.60% / 3.89% -6.57% / 3.90% -6.59% / 3.85% -6.60% / 3.97%
## 10: median bias MUE -0.15% / 0.81% -0.16% / 0.50% -0.12% / 0.80% -0.12% / 0.53% -0.13% / 0.53%
createTableResSim(res3stage.ar5binding, xtable = TRUE)
## *** ar 10 binding
res3stage.ar10binding <- res3stage[method.char != "method 3 fixC" & missing==TRUE & binding==TRUE & ar==10 & !is.na(decision),.SD,.SDcols=keep.col]
res3stage.ar10binding[,.N,by=c("method.char","type","hypo")][type=="interim",unique(N)]
## [1] 10000
table3stage.ar10binding <- createTableResSim(res3stage.ar10binding, xtable = FALSE)
table3stage.ar10binding
## statistic method 1 method 1 fixC method 2 method 2 fixC method 3
## 1: type 1 error 2.60% 2.47% 2.60% 2.50% 2.49%
## 2: power 74.51% 73.84% 74.51% 73.91% 74.01%
## 3: CI [NA,NA] 0.12% 0.20% 0.12% 0.20% 0.23%
## 4: coverage 95.07% 96.87% 95.07% 96.91% 96.12%
## 5: reversal 0.25%/0.08% 0.04%/0.54% 0.25%/0.08% 0.04%/0.50% 0/0.73%
## 6: abnormal 0.67%/0 0/0 0.67%/0 0/0 0/0.03%
## 7: mean bias LMME -0.035 / 0.021 -0.035 / 0.021 -0.035 / 0.021 -0.034 / 0.022 -0.034 / 0.023
## 8: mean bias MUE -0.027 / 0.023 -0.028 / 0.008 -0.027 / 0.023 -0.028 / 0.008 -0.028 / 0.014
## 9: median bias LMME -3.89% / 2.84% -3.91% / 2.85% -3.88% / 2.84% -3.80% / 2.85% -3.80% / 3.07%
## 10: median bias MUE -0.01% / 0.72% -0.05% / 0.13% 0.02% / 0.70% 0.08% / 0.19% 0.13% / 0.25%
createTableResSim(res3stage.ar10binding, xtable = TRUE)
## *** ar 5 non-binding
res3stage.ar5nonbinding <- res3stage[method.char != "method 3 fixC" & missing==TRUE & binding==FALSE & ar==5 & !is.na(decision),.SD,.SDcols=keep.col]
res3stage.ar5nonbinding[,.N,by=c("method.char","type","hypo")][type=="interim",unique(N)]
## [1] 10000
table3stage.ar5nonbinding <- createTableResSim(res3stage.ar5nonbinding, xtable = FALSE)
table3stage.ar5nonbinding
## statistic method 1 method 1 fixC method 2 method 2 fixC method 3
## 1: type 1 error 2.53% 2.43% 2.53% 2.44% 2.44%
## 2: power 74.61% 74.33% 74.59% 74.37% 74.43%
## 3: CI [NA,NA] 4.85% 5.13% 4.86% 5.00% 5.09%
## 4: coverage 95.87% 98.24% 95.89% 98.24% 97.00%
## 5: reversal 0.01%/0.01% 0.01%/0.29% 0.01%/0.01% 0.01%/0.28% 0/0.32%
## 6: abnormal 0.28%/0 0/0 0.28%/0 0/0 0/0.01%
## 7: mean bias LMME 0.004 / 0.065 0.003 / 0.065 0.004 / 0.065 0.003 / 0.065 0.003 / 0.066
## 8: mean bias MUE 0.004 / 0.057 0.003 / 0.038 0.004 / 0.057 0.003 / 0.037 0.004 / 0.049
## 9: median bias LMME -0.35% / 5.44% -0.40% / 5.54% -0.35% / 5.44% -0.40% / 5.43% -0.41% / 5.62%
## 10: median bias MUE -0.34% / 2.55% -0.39% / 2.41% -0.35% / 2.55% -0.39% / 2.35% -0.39% / 2.52%
createTableResSim(res3stage.ar5nonbinding, xtable = TRUE)
## *** ar 10 non-binding
res3stage.ar10nonbinding <- res3stage[method.char != "method 3 fixC" & missing==TRUE & binding==FALSE & ar==10 & !is.na(decision),.SD,.SDcols=keep.col]
res3stage.ar10nonbinding[,.N,by=c("method.char","type","hypo")][type=="interim",unique(N)]
## [1] 10000
table3stage.ar10nonbinding <- createTableResSim(res3stage.ar10nonbinding, xtable = FALSE)
table3stage.ar10nonbinding
## statistic method 1 method 1 fixC method 2 method 2 fixC method 3
## 1: type 1 error 2.49% 2.37% 2.49% 2.37% 2.42%
## 2: power 74.71% 74.17% 74.71% 74.25% 74.45%
## 3: CI [NA,NA] 4.78% 5.25% 4.79% 4.88% 4.99%
## 4: coverage 96.02% 97.96% 96.01% 97.95% 97.41%
## 5: reversal 0.21%/0.09% 0.05%/0.47% 0.21%/0.09% 0.04%/0.46% 0/0.57%
## 6: abnormal 0.54%/0 0/0 0.54%/0 0/0 0/0.03%
## 7: mean bias LMME 0.002 / 0.046 0.002 / 0.046 0.002 / 0.046 0.002 / 0.045 0.002 / 0.046
## 8: mean bias MUE 0.003 / 0.051 0.002 / 0.037 0.003 / 0.051 0.002 / 0.036 0.003 / 0.042
## 9: median bias LMME -0.40% / 4.29% -0.45% / 4.60% -0.40% / 4.29% -0.45% / 4.37% -0.47% / 4.58%
## 10: median bias MUE -0.40% / 2.60% -0.45% / 2.33% -0.40% / 2.59% -0.45% / 2.09% -0.39% / 2.21%
createTableResSim(res3stage.ar10nonbinding, xtable = TRUE)
##----------------------------------------------------------------------
### table1.R ends here
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.