Updated simirt.R

main
Corentin Choisy 1 year ago
parent a82848938e
commit b97d14840b

@ -1,5 +1,5 @@
simIRT <- function(NBOBS=2000,DIM,MU,COV,COVM,DIS,DIF,PMI,PMAX,ACC,CLEAR,STORE,REP,PREF,DRAW,
DRAWALL,ICC,GR=0,RAND,DEL=0,RSM1,RSM2,THR,TIT,PCM,ID,GENP,GENI) {
simIRT <- function(NBOBS=2000,DIM=NULL,MU=NULL,COV=NULL,COVM=NULL,DIS=NULL,DIF=NULL,PMIN=NULL,PMAX=NULL,ACC=NULL,CLEAR=NULL,STORE=NULL,REP=NULL,PREF=NULL,DRAW=NULL,
TYPEDIF=NULL,DRAWALL=NULL,ICC,GR=0,RAND=NULL,DELTAGR=0,RSM1=NULL,RSM2=NULL,THR=NULL,TIT=NULL,PCM=NULL,ID=NULL,GENP=NULL,GENI=NULL) {
if (GR < 0 | GR > 1) {
stop('Error 198: The GR option defines a probability. The values defined by this option must be greater (or equal) to 0 and lesser (or equal) to 1.')
@ -26,7 +26,7 @@ if (!is.null(DIM)) {
if (!is.null(DIF)) {
nbdiff <- length(DIF)
tmp <- DIF[1]
tmp <- TYPEDIF
if (tmp=='gauss' | tmp=='uniform') {
typediff <- tmp
}
@ -44,9 +44,171 @@ if (!is.null(DIM)) {
nbdiff <- length(DIF)*3
}
}
else if (is.null(DIM)) {
if (is.null(DIF) & is.null(PCM)) {
stop('Error 198: You must indicate the number of items to simulate in the DIM, DIF or PCM options.')
}
else if (!is.null(COVM)) {
nbrowcovm <- nrow(COVM)
if (nbrowcovm>1) {
stop('Error 198: You affected dimensions with COVM, but you did not affect each item to a dimension with the DIM option. Please correct DIM.')
}
}
else if (!is.null(PCM)) {
nbitems <- nrow(PCM)
DIM <- c(1)
}
else {
nbdiff <- length(DIF)
nbitems <- nbdiff
DIM <- c(1)
}
}
if ( (GR!=0 | DELTAGR!=0) & any(DIM != 1) & length(DIM)==1 ) {
stop('The GR and DELTAGR options are available only with unidimensional simulated data.')
}
if (is.null(PREF)) {
PREF <- 'item'
}
nbprefix <- length(PREF)
if (nbprefix!=length(DIM) & nbprefix!=1) { # ici, DIM est un vecteur contenant les nombres d'items par dim, mais est juste le nbre de dim dans stata
stop('Error 198: The PREF option is incorrect because the number of prefixes is different from the number of dimensions set in DIM. Please correct.')
}
if (nbprefix!=length(DIM)) {
alphab <- c('A','B','C','D','E','F','G','H','I','J','K','L','M','N','O','P','Q','R','S','T','U','V','W','X','Y','Z')
prefix <- paste0(prefix,alphab[seq(1,length(DIM))])
}
if (is.null(COVM)) {
nbcov <- length(COV)
if (length(DIM)==1) {
if (is.null(COV)) {
COV <- 1
}
if (nbcov>1) {
stop('Error 198: You simulate only one dimension, you must indicate only the variance of the simulated latent trait in the COV argument.')
}
if (COV <0) {
stop('Error 198: The variance of the latent trait cannot be negative. Please correct the COV option.')
}
covmatrix2 <- COV # cov doit être une matrice pré formatée
}
else if (length(DIM)==2) {
if (nbcov!=4 & nbcov>0) {
stop("Error 198: You simulate 2 dimensions. The COV option must be a 2x2 covariance matrix.")
}
else if (nbcov==0) {
if (length(DIM)==1) {
COV <- 1
}
else if (length(DIM)==2) {
COV <- matrix(c(1,0,0,1),nrow=2,byrow = T)
}
nbcov <- length(COV)
}
if (nbcov==4) {
cov1 <- COV[1,1]
cov2 <- COV[2,2]
cov3 <- COV[1,2]
rho <- cov3/sqrt(cov1*cov2)
if (cov1<0 | cov2<0 | rho< -1 | rho>1) {
stop('Error 198: The covariance matrix in COV is not correct. Please correct it.')
}
}
covmatrix2 <- COV
}
COVM <- COV
}
nbmu <- length(MU)
if (nbmu!=length(DIM) & nbmu!=0) {
stop('Error 198: You must indicate as many values in MU as the number of dimensions in DIM.')
}
nbdisc <- length(DIS)
if (nbdisc!=nbitems & nbdisc!=0) {
stop('Error 198: You must indicate as many values in DISC as the number of items in DIM and DIF.')
}
nbpmin <- length(PMIN)
if (nbpmin!=nbitems & nbpmin!=0) {
stop('Error 198: You must indicate as many values in PMIN as the number of items in DIM and DIF.')
}
nbpmax <- length(PMAX)
if (nbpmax!=nbitems & nbpmax!=0) {
stop('Error 198: You must indicate as many values in PMAX as the number of items in DIM and DIF.')
}
nbacc <- length(ACC)
if (nbacc != nbitems & nbacc!=0) {
stop('Error 198: You must indicate as many values in ACC as the number of items in DIM and DIF.')
}
if (!is.null(THR) & any(!is.null(c(DIS,PMIN,PMAX,ACC)))) {
stop('Error 198: If you use the THR option, you cannot use DIS, PMIN, PMAX or ACC.')
}
if (any(!is.null(c(RSM1,RSM2))) & any(!is.null(c(DIS,PMIN,PMAX,ACC)))) {
stop('Error 198: If you use the RSM1 and/or RSM2 option, you cannot use DIS, PMIN, PMAX or ACC.')
}
if (!is.null(PCM) & any(!is.null(c(DIS,PMIN,PMAX,ACC)))) {
stop('Error 198: If you use the PCM option, you cannot use DIS, PMIN, PMAX or ACC.')
}
if (any(!is.null(c(RSM1,RSM2))) & !is.null(PCM)) {
stop('Error 198: If you use the RSM1 and/or RSM2 option, you cannot use PCM.')
}
if (!is.null(RSM2) & length(DIM)==1) {
stop('Error 198: You cannot use RSM2 if you simulate only one dimension.')
}
if (is.null(ID)) {
ID <- 'ID'
}
##### Paramètres
hour <- as.numeric(substr(Sys.time(),12,13))
minu <- as.numeric(substr(Sys.time(),15,16))
sec <- as.numeric(substr(Sys.time(),18,19))
jour <- as.numeric(substr(Sys.Date(),9,10))
s <- 256484+1000000*sec+10000*minu+100*hour+jour
set.seed(s)
while(s>2^31-1) {
s <- s/231
}
set.seed(s)
if (typediff=='uniform') {
if (nbdiff %/% 2*length(DIM)==1) {
min <- DIF[(1:length(DIM)-1)*2+2]
max <- DIF[(1:length(DIM)-1)*2+3]
}
else if (nbdiff==1) {
min <- c(-2)
max <- c(2)
}
else {
stop('Your DIF option is incorrect. Please correct.')
}
for (d in seq(1,length(DIM))) {
for (i in seq(1,DIM[d])) {
DIF[i] <- min[d]+(max[d]-min[d])*i/(DIM[d]+1)
}
}
}
else if (typediff=='gauss') {
}

@ -151,6 +151,8 @@ if (`group'!=0|`deltagroup'!=0)&`dim'!=1 {
exit
}
if "`prefix'"=="" {
local prefix item
}
@ -172,6 +174,17 @@ else {
local prefix`d' `prefix'`tmp'
}
}
if "`covmatrix'"=="" {
tempname covmatrix2
local nbcov:word count `cov'
@ -191,6 +204,15 @@ if "`covmatrix'"=="" {
}
matrix `covmatrix2'=(`cov')
}
else if `dim'==2 {
if `nbcov'!=3&`nbcov'>0 {
di in red "You simulate two dimensions. You must indicate exactly 3 values in the {hi:cov} option (Variance of the first simulated latent trait, Variance of the second simulated latent trait, Covariance between the two simulated latent traits)."
@ -222,6 +244,13 @@ if "`covmatrix'"=="" {
local covmatrix `covmatrix2'
}
local nbmu:word count `mu'
if `nbmu'!=`dim'&`nbmu'!=0 {
di in red "You must indicate as many values in the {hi:mu} option as the number of dimension(s) (`dim')"
@ -316,6 +345,10 @@ while $seed>2^31-1 {
qui set seed $seed
if "`typediff'"=="uniform" {
if `nbdiff'==`=`dim'*2+1' {
local min`d':word `=(`d'-1)*2+2' of `diff'
@ -336,6 +369,14 @@ if "`typediff'"=="uniform" {
}
}
}
else if "`typediff'"=="gauss" {
if `nbdiff'==`=`dim'*2+1' {
forvalues d=1/`dim' {
@ -363,6 +404,12 @@ else if "`typediff'"=="gauss" {
}
}
forvalues d=1/`dim' {
if "`rsm`d''"!="" {
local nbrsm`d':word count `rsm`d''

Loading…
Cancel
Save