corentinchoisy/simul_these
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Corrected errors in residuals analysis

Browse Source
This commit is contained in:
Corentin Choisy
2024-04-29 10:54:39 +02:00
parent 425ac25a7e
commit d65132a190
3 changed files with 177 additions and 280 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
RProject/Scripts/aggregation.R
View File

@ -531,6 +531,8 @@ compile_simulation2_rosali <- function(scenario) {
flexible.detect <- prop.perfect
moreflexible.detect <- sapply(1:1000,function(x) ifelse(sum(!is.na(s[x,paste0("dif_detect_",1:4)]))!=0,s[x,"real_dif_1"]%in%c(s[x,paste0("dif_detect_",1:4)]),0) )
moreflexible.detect <- mean(moreflexible.detect)
percent.detect <- sapply(1:1000,function(x) sum(s[x,paste0("dif_detect_",1:4)]%in%c(s[x,c("real_dif_1")]))/1)
percent.detect <- mean(percent.detect)
}
if (nb.dif.true==2 & unique(b$J)==4 & unique(b$M)==2) {
perfect.detection <- sapply(1:1000,function(x) ifelse(sum(!is.na(s[x,paste0("dif_detect_",1:4)]))==2,s[x,paste0('dif_detect_',which(sapply(paste0("dif_detect_",1:4),function(y) !is.na(s[x,y])))[1])]%in%c(s[x,c("real_dif_1","real_dif_2")]) & s[x,paste0('dif_detect_',which(sapply(paste0("dif_detect_",1:4),function(y) !is.na(s[x,y])))[2])]%in%c(s[x,c("real_dif_1","real_dif_2")])
@ -540,6 +542,9 @@ compile_simulation2_rosali <- function(scenario) {
moreflexible.detect <- sapply(1:1000,function(x) ifelse(sum(!is.na(s[x,paste0("dif_detect_",1:4)]))!=0,s[x,"real_dif_1"]%in%c(s[x,paste0("dif_detect_",1:4)]) &
s[x,"real_dif_2"]%in%c(s[x,paste0("dif_detect_",1:4)]),0) )
moreflexible.detect <- mean(moreflexible.detect)
percent.detect <- sapply(1:1000,function(x) sum(s[x,paste0("dif_detect_",1:4)]%in%c(s[x,c("real_dif_1","real_dif_2")]))/2)
percent.detect <- mean(percent.detect)
}
if (nb.dif.true==2 & unique(b$J)==7 & unique(b$M)==2) {
perfect.detection <- sapply(1:1000,function(x) ifelse(sum(!is.na(s[x,paste0("dif_detect_",1:7)]))==2,s[x,paste0('dif_detect_',which(sapply(paste0("dif_detect_",1:7),function(y) !is.na(s[x,y])))[1])]%in%c(s[x,c("real_dif_1","real_dif_2")]) & s[x,paste0('dif_detect_',which(sapply(paste0("dif_detect_",1:7),function(y) !is.na(s[x,y])))[2])]%in%c(s[x,c("real_dif_1","real_dif_2")])
@ -549,6 +554,8 @@ compile_simulation2_rosali <- function(scenario) {
moreflexible.detect <- sapply(1:1000,function(x) ifelse(sum(!is.na(s[x,paste0("dif_detect_",1:7)]))!=0,s[x,"real_dif_1"]%in%c(s[x,paste0("dif_detect_",1:7)]) &
s[x,"real_dif_2"]%in%c(s[x,paste0("dif_detect_",1:7)]),0) )
moreflexible.detect <- mean(moreflexible.detect)
percent.detect <- sapply(1:1000,function(x) sum(s[x,paste0("dif_detect_",1:7)]%in%c(s[x,c("real_dif_1","real_dif_2")]))/2)
percent.detect <- mean(percent.detect)
}
if (nb.dif.true==3 & unique(b$J)==7 & unique(b$M)==2) {
perfect.detection <- sapply(1:1000,function(x) ifelse(sum(!is.na(s[x,paste0("dif_detect_",1:7)]))==3,s[x,paste0('dif_detect_',which(sapply(paste0("dif_detect_",1:7),function(y) !is.na(s[x,y])))[1])]%in%c(s[x,c("real_dif_1","real_dif_2","real_dif_3")]) & s[x,paste0('dif_detect_',which(sapply(paste0("dif_detect_",1:7),function(y) !is.na(s[x,y])))[2])]%in%c(s[x,c("real_dif_1","real_dif_2","real_dif_3")]) & s[x,paste0('dif_detect_',which(sapply(paste0("dif_detect_",1:7),function(y) !is.na(s[x,y])))[3])]%in%c(s[x,c("real_dif_1","real_dif_2","real_dif_3")])
@ -558,6 +565,8 @@ compile_simulation2_rosali <- function(scenario) {
moreflexible.detect <- sapply(1:1000,function(x) ifelse(sum(!is.na(s[x,paste0("dif_detect_",1:7)]))!=0,s[x,"real_dif_1"]%in%c(s[x,paste0("dif_detect_",1:7)]) &
s[x,"real_dif_2"]%in%c(s[x,paste0("dif_detect_",1:7)]) & s[x,"real_dif_3"]%in%c(s[x,paste0("dif_detect_",1:7)]),0) )
moreflexible.detect <- mean(moreflexible.detect)
percent.detect <- sapply(1:1000,function(x) sum(s[x,paste0("dif_detect_",1:7)]%in%c(s[x,c("real_dif_1","real_dif_2","real_dif_3")]))/3)
percent.detect <- mean(percent.detect)
}
z <- data.frame(m.beta=mean(s$beta),
se.empirical.beta=sd(s$beta),
@ -571,7 +580,8 @@ compile_simulation2_rosali <- function(scenario) {
dif.detected=dif.d,
prop.perfect=prop.perfect,
flexible.detect=flexible.detect,
moreflexible.detect=moreflexible.detect
moreflexible.detect=moreflexible.detect,
percent.detect=ifelse(name%%2==0,NA,percent.detect)
)
d <- cbind(b,a,z)
d$prop.

160
RProject/Scripts/functions/resali.R
View File

@ -18,6 +18,8 @@ resali <- function(df=NULL,items=NULL,group=NULL,verbose=T) {
while (length(unique(quantile(dat$score,seq(0,1,1/nqt))))!=nqt+1) {
nqt <- nqt-1
}
# ITEM POLYTOMIQUE
if (max(resp)>1) {
dat$score_q5 <- cut(dat$score,unique(quantile(dat$score,seq(0,1,1/nqt))),labels=1:nqt,include.lowest=T)
res.anova <- rep(NA,nbitems)
pval <- rep(NA,nbitems)
@ -89,67 +91,15 @@ resali <- function(df=NULL,items=NULL,group=NULL,verbose=T) {
}
return(NULL)
}
}
resali_BF2 <- function(df=NULL,items=NULL,group=NULL,verbose=T) {
if (verbose) {
cat('-----------------------------------------------------------\n')
cat('COMPUTING INITIAL PCM\n')
cat('-----------------------------------------------------------\n')
}
nbitems <- length(items)
nbitems_o <- nbitems
items_n <- paste0('item',items)
resp <- df[,items_n]
grp <- df[,group]
pcm_initial <- TAM::tam.mml(resp=resp,Y=grp,irtmodel = "PCM",est.variance = T,verbose=F)
dat <- df
dat$score <- rowSums(dat[,items_n])
nqt <- ifelse(length(unique(quantile(dat$score,seq(0,1,0.2))))==6,5,length(unique(quantile(dat$score,seq(0,1,0.2))))-1)
dat$score_q5 <- cut(dat$score,unique(quantile(dat$score,seq(0,1,1/nqt))),labels=1:nqt,include.lowest=T)
res.anova <- rep(NA,nbitems)
pval <- rep(NA,nbitems)
fval <- rep(NA,nbitems)
for (i in items) {
dat[,paste0('res_',i)] <- rowMeans(predict(pcm_initial)$stand.resid[,,i])
res.anova[i] <- summary(aov(dat[,paste0('res_',i)]~TT*score_q5,data=dat))
pval[i] <- res.anova[[i]][1,"Pr(>F)"]
fval[i] <- res.anova[[i]][1,'F value']
}
if (verbose) {
cat('DONE\n')
cat('-----------------------------------------------------------\n')
}
res.items <- c()
res.uniform <- c()
k <- 1
while(any(pval<0.05/(nbitems_o-k+1))) {
k <- k+1
if (verbose) {
cat(paste('COMPUTING STEP',k,'\n'))
cat('-----------------------------------------------------------\n')
}
res.item <- gsub("[a-z]", "",colnames(resp)[which.max(fval)])
res.items <- c(res.items,res.item)
res.uni <- res.anova[[which.max(fval)]][3,"Pr(>F)"]>0.05
res.uniform <- c(res.uniform,res.uni)
items_n <- c(items_n[items_n!=paste0('item',res.item)],paste0("item",res.item,c("noTT","TT")))
dat[,paste0("item",res.item,'TT')] <- dat[dat$TT==1,paste0('item',res.item)]
dat[,paste0("item",res.item,'noTT')] <- dat[dat$TT==0,paste0('item',res.item)]
resp <- dat[,items_n]
grp <- dat[,group]
pcm_while <- TAM::tam.mml(resp=resp,Y=grp,irtmodel = "PCM",est.variance = T,verbose=F)
nbitems <- length(items_n)
# ITEM DICHOTOMIQUE
} else {
res.anova <- rep(NA,nbitems)
pval <- rep(NA,nbitems)
fval <- rep(NA,nbitems)
for (i in 1:nbitems) {
dat[,paste0('res_',i)] <- rowMeans(predict(pcm_while)$stand.resid[,,i])
res.anova[i] <- summary(aov(dat[,paste0('res_',i)]~TT*score_q5,data=dat))
for (i in items) {
dat[,paste0('res_',i)] <- IRT.residuals(pcm_initial)$stand_residuals[,i]
res.anova[i] <- summary(aov(dat[,paste0('res_',i)]~TT,data=dat))
pval[i] <- res.anova[[i]][1,"Pr(>F)"]
fval[i] <- res.anova[[i]][1,'F value']
}
@ -157,72 +107,9 @@ resali_BF2 <- function(df=NULL,items=NULL,group=NULL,verbose=T) {
cat('DONE\n')
cat('-----------------------------------------------------------\n')
}
}
if (verbose) {
cat("DETECTED DIF ITEMS\n")
cat('-----------------------------------------------------------\n')
}
if (length(res.items>0)) {
results <- data.frame(dif.items=res.items,
uniform=1*res.uniform)
return(results)
}
else {
if (verbose) {
cat("No DIF was detected\n")
}
return(NULL)
}
}
resali_LRT <- function(df=NULL,items=NULL,group=NULL,verbose=T) {
if (verbose) {
cat('-----------------------------------------------------------\n')
cat('COMPUTING INITIAL PCM\n')
cat('-----------------------------------------------------------\n')
}
nbitems <- length(items)
nbitems_o <- nbitems
items_n <- paste0('item',items)
resp <- df[,items_n]
grp <- df[,group]
items_n_alt <- paste0(items_n,c("_TT","_noTT"))
for (i in items_n) {
df[df$TT==0,paste0(i,"_noTT")] <- df[df$TT==0,i]
df[df$TT==1,paste0(i,"_TT")] <- df[df$TT==1,i]
}
resp_alt <- df[,items_n_alt]
pcm_initial <- TAM::tam.mml(resp=resp,Y=grp,irtmodel = "PCM",est.variance = T,verbose=F)
pcm_alt <- TAM::tam.mml(resp=resp_alt,Y=grp,irtmodel = "PCM",est.variance = T,verbose=F)
dat <- df
dat$score <- rowSums(dat[,items_n])
nqt <- ifelse(length(unique(quantile(dat$score,seq(0,1,0.2))))==6,5,length(unique(quantile(dat$score,seq(0,1,0.2))))-1)
dat$score_q5 <- cut(dat$score,unique(quantile(dat$score,seq(0,1,1/nqt))),labels=1:nqt,include.lowest=T)
res.anova <- rep(NA,nbitems)
pval <- rep(NA,nbitems)
fval <- rep(NA,nbitems)
for (i in items) {
dat[,paste0('res_',i)] <- rowMeans(predict(pcm_initial)$stand.resid[,,i])
res.anova[i] <- summary(aov(dat[,paste0('res_',i)]~TT*score_q5,data=dat))
pval[i] <- res.anova[[i]][1,"Pr(>F)"]
fval[i] <- res.anova[[i]][1,'F value']
}
if (verbose) {
cat('DONE\n')
cat('-----------------------------------------------------------\n')
}
res.items <- c()
res.uniform <- c()
k <- 1
if (anova(pcm_initial,pcm_alt)$p[1]<0.05) {
while(any(pval<0.05/nbitems_o)) {
res.items <- c()
k <- 1
while(any(pval<0.05/(nbitems_o))) {
k <- k+1
if (verbose) {
cat(paste('COMPUTING STEP',k,'\n'))
@ -230,8 +117,6 @@ resali_LRT <- function(df=NULL,items=NULL,group=NULL,verbose=T) {
}
res.item <- gsub("[a-z]", "",colnames(resp)[which.max(fval)])
res.items <- c(res.items,res.item)
res.uni <- res.anova[[which.max(fval)]][3,"Pr(>F)"]>0.05
res.uniform <- c(res.uniform,res.uni)
items_n <- c(items_n[items_n!=paste0('item',res.item)],paste0("item",res.item,c("noTT","TT")))
dat[,paste0("item",res.item,'TT')] <- dat[dat$TT==1,paste0('item',res.item)]
dat[,paste0("item",res.item,'noTT')] <- dat[dat$TT==0,paste0('item',res.item)]
@ -243,11 +128,19 @@ resali_LRT <- function(df=NULL,items=NULL,group=NULL,verbose=T) {
pval <- rep(NA,nbitems)
fval <- rep(NA,nbitems)
for (i in 1:nbitems) {
dat[,paste0('res_',i)] <- rowMeans(predict(pcm_while)$stand.resid[,,i])
res.anova[i] <- summary(aov(dat[,paste0('res_',i)]~TT*score_q5,data=dat))
dat[,paste0('res_',i)] <- IRT.residuals(pcm_while)$stand_residuals[,i]
res.anova[i] <- summary(aov(dat[,paste0('res_',i)]~TT,data=dat))
pval[i] <- res.anova[[i]][1,"Pr(>F)"]
fval[i] <- res.anova[[i]][1,'F value']
}
zz <- 0
for (name_i in items_n) {
zz <- zz+1
if (grepl("TT",name_i)) {
pval[zz] <- 1
fval[zz] <- 0
}
}
if (verbose) {
cat('DONE\n')
cat('-----------------------------------------------------------\n')
@ -259,7 +152,7 @@ resali_LRT <- function(df=NULL,items=NULL,group=NULL,verbose=T) {
}
if (length(res.items>0)) {
results <- data.frame(dif.items=res.items,
uniform=1*res.uniform)
uniform=rep(1,length(res.items)))
return(results)
}
else {
@ -268,11 +161,6 @@ resali_LRT <- function(df=NULL,items=NULL,group=NULL,verbose=T) {
}
return(NULL)
}
}
else {
if (verbose) {
cat("No DIF was detected\n")
}
return(NULL)
}
}
}

285
RProject/Scripts/resali_generate_newdata.R
View File

@ -1,3 +1,145 @@
##############################################################################
#----------------------------------------------------------------------------#
################################### RESALI ###################################
#----------------------------------------------------------------------------#
##############################################################################
generate_resali <- function(scenario=NULL,grp=NULL) {
scen <- as.numeric(gsub("[A,B,C,D,E,F,G,_]","",substr(scenario,0,3)))
if (substr(scenario,start=nchar(scenario)-1,stop=nchar(scenario))=="50") {
N <- 50
}
if (substr(scenario,start=nchar(scenario)-2,stop=nchar(scenario))=="100") {
N <- 100
}
if (substr(scenario,start=nchar(scenario)-2,stop=nchar(scenario))=="200") {
N <- 200
}
if (substr(scenario,start=nchar(scenario)-2,stop=nchar(scenario))=="300") {
N <- 300
}
if (scen<5) {
dat <- read.csv(paste0('/home/corentin/Documents/These/Recherche/Simulations/Data/NoDIF/N',N,'/scenario_',scenario,'.csv'))
}
if (scen>=5) {
dat <- read.csv(paste0('/home/corentin/Documents/These/Recherche/Simulations/Data/DIF/N',N,'/scenario_',scenario,'.csv'))
}
if (scen%in%c(3,4,13:20)) {
res <- resali(df=dat[dat$replication==1,],items = seq(1,7),group=grp,verbose=FALSE)
df_res <- data.frame(dif.detect.1=ifelse(length(res$dif.items)>=1,res$dif.items[1],NA),
dif.detect.2=ifelse(length(res$dif.items)>=2,res$dif.items[2],NA),
dif.detect.3=ifelse(length(res$dif.items)>=3,res$dif.items[3],NA),
dif.detect.4=ifelse(length(res$dif.items)>=4,res$dif.items[4],NA),
dif.detect.5=ifelse(length(res$dif.items)>=5,res$dif.items[5],NA),
dif.detect.6=ifelse(length(res$dif.items)>=6,res$dif.items[6],NA),
dif.detect.7=ifelse(length(res$dif.items)>=7,res$dif.items[7],NA),
dif.detect.unif.1=ifelse(length(res$uniform)>=1,res$uniform[1],NA),
dif.detect.unif.2=ifelse(length(res$uniform)>=2,res$uniform[2],NA),
dif.detect.unif.3=ifelse(length(res$uniform)>=3,res$uniform[3],NA),
dif.detect.unif.4=ifelse(length(res$uniform)>=4,res$uniform[4],NA),
dif.detect.unif.5=ifelse(length(res$uniform)>=5,res$uniform[5],NA),
dif.detect.unif.6=ifelse(length(res$uniform)>=6,res$uniform[6],NA),
dif.detect.unif.7=ifelse(length(res$uniform)>=7,res$uniform[7],NA),
N=N,
nbdif=ifelse(scen<=4,0,ifelse(scen<=16,2,3)),
true.dif.1=ifelse(scen<=4,NA,unique(dat[dat$replication==1,]$dif1)),
true.dif.2=ifelse(scen<=4,NA,unique(dat[dat$replication==1,]$dif2)),
true.dif.3=ifelse(scen<=16,NA,unique(dat[dat$replication==1,]$dif3))
)
for (k in 2:1000) {
if (k%%100==0) {
cat(paste0('N=',k,'/1000\n'))
}
res <- resali(df=dat[dat$replication==k,],items = seq(1,7),group=grp,verbose=FALSE)
df_res2 <- data.frame(dif.detect.1=ifelse(length(res$dif.items)>=1,res$dif.items[1],NA),
dif.detect.2=ifelse(length(res$dif.items)>=2,res$dif.items[2],NA),
dif.detect.3=ifelse(length(res$dif.items)>=3,res$dif.items[3],NA),
dif.detect.4=ifelse(length(res$dif.items)>=4,res$dif.items[4],NA),
dif.detect.5=ifelse(length(res$dif.items)>=5,res$dif.items[5],NA),
dif.detect.6=ifelse(length(res$dif.items)>=6,res$dif.items[6],NA),
dif.detect.7=ifelse(length(res$dif.items)>=7,res$dif.items[7],NA),
dif.detect.unif.1=ifelse(length(res$uniform)>=1,res$uniform[1],NA),
dif.detect.unif.2=ifelse(length(res$uniform)>=2,res$uniform[2],NA),
dif.detect.unif.3=ifelse(length(res$uniform)>=3,res$uniform[3],NA),
dif.detect.unif.4=ifelse(length(res$uniform)>=4,res$uniform[4],NA),
dif.detect.unif.5=ifelse(length(res$uniform)>=5,res$uniform[5],NA),
dif.detect.unif.6=ifelse(length(res$uniform)>=6,res$uniform[6],NA),
dif.detect.unif.7=ifelse(length(res$uniform)>=7,res$uniform[7],NA),
N=N,
nbdif=ifelse(scen<=4,0,ifelse(scen<=16,2,3)),
true.dif.1=ifelse(scen<=4,NA,unique(dat[dat$replication==k,]$dif1)),
true.dif.2=ifelse(scen<=4,NA,unique(dat[dat$replication==k,]$dif2)),
true.dif.3=ifelse(scen<=16,NA,unique(dat[dat$replication==k,]$dif3)))
df_res <- rbind(df_res,df_res2)
}
}
else if (scen%in%c(1,2,5:12)) {
res <- resali(df=dat[dat$replication==1,],items = seq(1,4),group=grp,verbose=FALSE)
df_res <- data.frame(dif.detect.1=ifelse(length(res$dif.items)>=1,res$dif.items[1],NA),
dif.detect.2=ifelse(length(res$dif.items)>=2,res$dif.items[2],NA),
dif.detect.3=ifelse(length(res$dif.items)>=3,res$dif.items[3],NA),
dif.detect.4=ifelse(length(res$dif.items)>=4,res$dif.items[4],NA),
dif.detect.unif.1=ifelse(length(res$uniform)>=1,res$uniform[1],NA),
dif.detect.unif.2=ifelse(length(res$uniform)>=2,res$uniform[2],NA),
dif.detect.unif.3=ifelse(length(res$uniform)>=3,res$uniform[3],NA),
dif.detect.unif.4=ifelse(length(res$uniform)>=4,res$uniform[4],NA),
N=N,
nbdif=ifelse(scen<=4,0,ifelse(scen<=8,1,2)),
true.dif.1=ifelse(scen<=4,NA,unique(dat[dat$replication==1,]$dif1)),
true.dif.2=ifelse(scen<=8,NA,unique(dat[dat$replication==1,]$dif2))
)
for (k in 2:1000) {
if (k%%100==0) {
cat(paste0('N=',k,'/1000\n'))
}
res <- resali(df=dat[dat$replication==k,],items = seq(1,4),group=grp,verbose=FALSE)
df_res2 <- data.frame(dif.detect.1=ifelse(length(res$dif.items)>=1,res$dif.items[1],NA),
dif.detect.2=ifelse(length(res$dif.items)>=2,res$dif.items[2],NA),
dif.detect.3=ifelse(length(res$dif.items)>=3,res$dif.items[3],NA),
dif.detect.4=ifelse(length(res$dif.items)>=4,res$dif.items[4],NA),
dif.detect.unif.1=ifelse(length(res$uniform)>=1,res$uniform[1],NA),
dif.detect.unif.2=ifelse(length(res$uniform)>=2,res$uniform[2],NA),
dif.detect.unif.3=ifelse(length(res$uniform)>=3,res$uniform[3],NA),
dif.detect.unif.4=ifelse(length(res$uniform)>=4,res$uniform[4],NA),
N=N,
nbdif=ifelse(scen<=4,0,ifelse(scen<=8,1,2)),
true.dif.1=ifelse(scen<=4,NA,unique(dat[dat$replication==k,]$dif1)),
true.dif.2=ifelse(scen<=8,NA,unique(dat[dat$replication==k,]$dif2)))
df_res <- rbind(df_res,df_res2)
}
}
return(df_res)
}
results <- c(sapply(1:4,function(x) paste0(x,c('A','B','C','D','E'))),sapply(5:9,function(x) paste0(x,c('A','B','C','D','E','F','G'))))
results2 <- c(sapply(10:20,function(x) paste0(x,c('A','B','C','D','E','F','G'))))
results <- c(sapply(c(50,100,200,300),function(x) paste0(results,'_',x)))
results2 <- c(sapply(c(50,100,200,300),function(x) paste0(results2,'_',x)))
results <- sort(results)
results2 <- sort(results2)
results <- c(results,results2)
for (r in results) {
cat(paste0(r,"\n"))
cat(paste0("-------------------------------------------","\n"))
write.csv(generate_resali(r,"TT"),paste0("/home/corentin/Documents/These/Recherche/Simulations/Analysis/RESALI/Detection/",r,".csv"))
cat(paste0("-------------------------------------------","\n"))
}
##############################################################################
#----------------------------------------------------------------------------#
################################### NEWDATA ##################################
#----------------------------------------------------------------------------#
##############################################################################
## Liste des scenarios
results <- c(sapply(1:4,function(x) paste0(x,c('A','B','C','D','E'))),sapply(5:9,function(x) paste0(x,c('A','B','C','D','E','F','G'))))
@ -135,146 +277,3 @@ for (r in results) {
write.csv(analyse,paste0("/home/corentin/Documents/These/Recherche/Simulations/Analysis/RESALI/Detection_data/",r,".csv"))
}
##############################################################################
#----------------------------------------------------------------------------#
################################### RESALI ###################################
#----------------------------------------------------------------------------#
##############################################################################
generate_resali <- function(scenario=NULL,grp=NULL) {
scen <- as.numeric(gsub("[A,B,C,D,E,F,G,_]","",substr(scenario,0,3)))
if (substr(scenario,start=nchar(scenario)-1,stop=nchar(scenario))=="50") {
N <- 50
}
if (substr(scenario,start=nchar(scenario)-2,stop=nchar(scenario))=="100") {
N <- 100
}
if (substr(scenario,start=nchar(scenario)-2,stop=nchar(scenario))=="200") {
N <- 200
}
if (substr(scenario,start=nchar(scenario)-2,stop=nchar(scenario))=="300") {
N <- 300
}
if (scen<5) {
dat <- read.csv(paste0('/home/corentin/Documents/These/Recherche/Simulations/Data/NoDIF/N',N,'/scenario_',scenario,'.csv'))
}
if (scen>=5) {
dat <- read.csv(paste0('/home/corentin/Documents/These/Recherche/Simulations/Data/DIF/N',N,'/scenario_',scenario,'.csv'))
}
if (scen%in%c(3,4,13:20)) {
res <- resali(df=dat[dat$replication==1,],items = seq(1,7),group=grp,verbose=FALSE)
df_res <- data.frame(dif.detect.1=ifelse(length(res$dif.items)>=1,res$dif.items[1],NA),
dif.detect.2=ifelse(length(res$dif.items)>=2,res$dif.items[2],NA),
dif.detect.3=ifelse(length(res$dif.items)>=3,res$dif.items[3],NA),
dif.detect.4=ifelse(length(res$dif.items)>=4,res$dif.items[4],NA),
dif.detect.5=ifelse(length(res$dif.items)>=5,res$dif.items[5],NA),
dif.detect.6=ifelse(length(res$dif.items)>=6,res$dif.items[6],NA),
dif.detect.7=ifelse(length(res$dif.items)>=7,res$dif.items[7],NA),
dif.detect.unif.1=ifelse(length(res$uniform)>=1,res$uniform[1],NA),
dif.detect.unif.2=ifelse(length(res$uniform)>=2,res$uniform[2],NA),
dif.detect.unif.3=ifelse(length(res$uniform)>=3,res$uniform[3],NA),
dif.detect.unif.4=ifelse(length(res$uniform)>=4,res$uniform[4],NA),
dif.detect.unif.5=ifelse(length(res$uniform)>=5,res$uniform[5],NA),
dif.detect.unif.6=ifelse(length(res$uniform)>=6,res$uniform[6],NA),
dif.detect.unif.7=ifelse(length(res$uniform)>=7,res$uniform[7],NA),
N=N,
nbdif=ifelse(scen<=4,0,ifelse(scen<=16,2,3)),
true.dif.1=ifelse(scen<=4,NA,unique(dat[dat$replication==1,]$dif1)),
true.dif.2=ifelse(scen<=4,NA,unique(dat[dat$replication==1,]$dif2)),
true.dif.3=ifelse(scen<=16,NA,unique(dat[dat$replication==1,]$dif3))
)
for (k in 2:1000) {
if (k%%100==0) {
cat(paste0('N=',k,'/1000\n'))
}
res <- resali(df=dat[dat$replication==k,],items = seq(1,7),group=grp,verbose=FALSE)
df_res2 <- data.frame(dif.detect.1=ifelse(length(res$dif.items)>=1,res$dif.items[1],NA),
dif.detect.2=ifelse(length(res$dif.items)>=2,res$dif.items[2],NA),
dif.detect.3=ifelse(length(res$dif.items)>=3,res$dif.items[3],NA),
dif.detect.4=ifelse(length(res$dif.items)>=4,res$dif.items[4],NA),
dif.detect.5=ifelse(length(res$dif.items)>=5,res$dif.items[5],NA),
dif.detect.6=ifelse(length(res$dif.items)>=6,res$dif.items[6],NA),
dif.detect.7=ifelse(length(res$dif.items)>=7,res$dif.items[7],NA),
dif.detect.unif.1=ifelse(length(res$uniform)>=1,res$uniform[1],NA),
dif.detect.unif.2=ifelse(length(res$uniform)>=2,res$uniform[2],NA),
dif.detect.unif.3=ifelse(length(res$uniform)>=3,res$uniform[3],NA),
dif.detect.unif.4=ifelse(length(res$uniform)>=4,res$uniform[4],NA),
dif.detect.unif.5=ifelse(length(res$uniform)>=5,res$uniform[5],NA),
dif.detect.unif.6=ifelse(length(res$uniform)>=6,res$uniform[6],NA),
dif.detect.unif.7=ifelse(length(res$uniform)>=7,res$uniform[7],NA),
N=N,
nbdif=ifelse(scen<=4,0,ifelse(scen<=16,2,3)),
true.dif.1=ifelse(scen<=4,NA,unique(dat[dat$replication==k,]$dif1)),
true.dif.2=ifelse(scen<=4,NA,unique(dat[dat$replication==k,]$dif2)),
true.dif.3=ifelse(scen<=16,NA,unique(dat[dat$replication==k,]$dif3)))
df_res <- rbind(df_res,df_res2)
}
}
else if (scen%in%c(1,2,5:12)) {
res <- resali(df=dat[dat$replication==1,],items = seq(1,4),group=grp,verbose=FALSE)
df_res <- data.frame(dif.detect.1=ifelse(length(res$dif.items)>=1,res$dif.items[1],NA),
dif.detect.2=ifelse(length(res$dif.items)>=2,res$dif.items[2],NA),
dif.detect.3=ifelse(length(res$dif.items)>=3,res$dif.items[3],NA),
dif.detect.4=ifelse(length(res$dif.items)>=4,res$dif.items[4],NA),
dif.detect.unif.1=ifelse(length(res$uniform)>=1,res$uniform[1],NA),
dif.detect.unif.2=ifelse(length(res$uniform)>=2,res$uniform[2],NA),
dif.detect.unif.3=ifelse(length(res$uniform)>=3,res$uniform[3],NA),
dif.detect.unif.4=ifelse(length(res$uniform)>=4,res$uniform[4],NA),
N=N,
nbdif=ifelse(scen<=4,0,ifelse(scen<=8,1,2)),
true.dif.1=ifelse(scen<=4,NA,unique(dat[dat$replication==1,]$dif1)),
true.dif.2=ifelse(scen<=8,NA,unique(dat[dat$replication==1,]$dif2))
)
for (k in 2:1000) {
if (k%%100==0) {
cat(paste0('N=',k,'/1000\n'))
}
res <- resali(df=dat[dat$replication==k,],items = seq(1,4),group=grp,verbose=FALSE)
df_res2 <- data.frame(dif.detect.1=ifelse(length(res$dif.items)>=1,res$dif.items[1],NA),
dif.detect.2=ifelse(length(res$dif.items)>=2,res$dif.items[2],NA),
dif.detect.3=ifelse(length(res$dif.items)>=3,res$dif.items[3],NA),
dif.detect.4=ifelse(length(res$dif.items)>=4,res$dif.items[4],NA),
dif.detect.unif.1=ifelse(length(res$uniform)>=1,res$uniform[1],NA),
dif.detect.unif.2=ifelse(length(res$uniform)>=2,res$uniform[2],NA),
dif.detect.unif.3=ifelse(length(res$uniform)>=3,res$uniform[3],NA),
dif.detect.unif.4=ifelse(length(res$uniform)>=4,res$uniform[4],NA),
N=N,
nbdif=ifelse(scen<=4,0,ifelse(scen<=8,1,2)),
true.dif.1=ifelse(scen<=4,NA,unique(dat[dat$replication==k,]$dif1)),
true.dif.2=ifelse(scen<=8,NA,unique(dat[dat$replication==k,]$dif2)))
df_res <- rbind(df_res,df_res2)
}
}
return(df_res)
}
results <- c(sapply(1:4,function(x) paste0(x,c('A','B','C','D','E'))),sapply(5:9,function(x) paste0(x,c('A','B','C','D','E','F','G'))))
results2 <- c(sapply(10:20,function(x) paste0(x,c('A','B','C','D','E','F','G'))))
results <- c(sapply(c(50,100,200,300),function(x) paste0(results,'_',x)))
results2 <- c(sapply(c(50,100,200,300),function(x) paste0(results2,'_',x)))
results <- sort(results)
results2 <- sort(results2)
results <- c(results,results2)
for (r in results) {
cat(paste0(r,"\n"))
cat(paste0("-------------------------------------------","\n"))
write.csv(generate_resali(r,"TT"),paste0("/home/corentin/Documents/These/Recherche/Simulations/Analysis/RESALI/Detection/",r,".csv"))
cat(paste0("-------------------------------------------","\n"))
}