Nothing
R/pn.mod.compare.R
In FlexParamCurve: Tools to Fit Flexible Parametric Curves
pn.mod.compare <-
structure(function # Compare All Possible Positive-Negative Richards \eqn{nlslist} Models
(x,
### a numeric vector of the primary predictor
y,
### a numeric vector of the response variable
grp,
### a factor of same length as x and y that distinguishes groups within
### the dataset
pn.options,
### required character value for name of list object populated with starting
### parameter estimates, fitting
### options and bounds or destination for modpar to write a new list (see Details)
forcemod = 0,
### optional numeric value to constrain model selection (see Details)
existing = FALSE,
### optional logical value specifying whether some of the relevant models
### have already been fitted
penaliz = "1/sqrt(n)",
### optional character value to determine how models are ranked (see Details)
taper.ends = 0.45,
mod.subset = c(NA),
Envir = .GlobalEnv,
...
) {
##description<< This function performs model selection for \code{\link{nlsList}} models fitted using
## \code{\link{SSposnegRichards}}.
##details<< First, whether parameter M should be fixed
## (see \code{\link{SSposnegRichards}}) is determined by fitting models 12 and 20 and comparing
## their perfomance using \code{\link{extraF}}.
## If model 12 provides superior performance (variable values of M) then 16 models that estimate M
## are run
## (models 1 through 16), otherwise the models with fixed M are fitted (models 21 through 36).
## Fitting these \code{\link{nlsList}} models can be time-consuming (2-4 hours using the dataset
## \code{\link{posneg.data}} that encompasses 100 individuals) and if several of the relevant
## models are already fitted the option existing=TRUE can be used to avoid refitting models that
## already exist (note that a model object in which no grouping levels were successfully
## parameterized will be refitted, as will objects that are not of class nlsList).
##
## Specifying forcemod=3 will force model selection to only consider fixed M models and setting
## forcemod=4 will force model selection to consider models with varying values of M only.
## If fitting both models
## 12 and 20 fails, fixed M models will be used by default.
##
## Models are ranked by modified pooled residual square error. By default residual standard error
## is divided by the square root of sample size. This exponentially penalizes models for which very few
## grouping levels (individuals) are successfully parameterized (the few individuals that are
## parameterized in these models are fit unsuprisingly well) using a function based on the relationship
## between standard error and sample size. However, different users may have different preferences
## and these can be specified in the argument penaliz (which residual
## standard error is multiplied by). This argument must be a character value
## that contains the character n (sample size) and must be a valid right hand side (RHS) of a formula:
## e.g. 1*(n), (n)^2. It cannot contain more than one n but could be a custom function, e.g. FUN(n).
if( anyNA(data.frame(x, y, grp)) == TRUE ) stop ("This function does not handle missing data values for x, y, or grp. Please subset the data, e.g. mydataframe[!is.na(mydataframe),], to remove them prior to function call")
options(warn = -1)
if(existing == FALSE) {
rm(list = ls(pattern = "richardsR", envir=FPCEnv), envir=FPCEnv)
rm(list = ls(pattern = "richardsR", envir=Envir), envir=Envir)
} else {
rm(list = ls(pattern = "richardsR", envir=FPCEnv), envir=FPCEnv)
get.mod(modelname = ls(Envir,pattern="richardsR"), from.envir = Envir, to.envir = FPCEnv, write.mod = TRUE, silent = TRUE)
print("#########################################################################################################")
print(paste("NOTE: existing is not set to false, existing models in the working environment ",substitute(Envir)," will be used during fits. To remove models manually, remove all files prefixed richardsR from working environment before running",sep=""))
print("#########################################################################################################")
}
alacarte <- FALSE
"%w/o%" <- function(x, y) x[!x %in% y] #-- x without y
if(length(mod.subset %w/o% NA) > 0) {
alacarte <- TRUE
}
pnoptnm <- as.character(pn.options[1])
checkpen <- try(unlist(strsplit(penaliz, "(n)")), silent = TRUE)
if (length(checkpen) != 2 | class(checkpen)[1] == "try-error") {
stop("penaliz parameter is ill defined: see ?pn.mod.compare")
} else {
checkpen <- try(eval(parse(text = sprintf("%s", paste(checkpen[1],
"1", checkpen[2], sep = "")))))
if (class(checkpen)[1] == "try-error")
stop("penaliz parameter is ill defined: see ?pn.mod.compare")
}
datamerg <- data.frame(x, y, grp)
userdata <- groupedData(y ~ x | grp, outer = ~grp, data = datamerg)
assign("userdata",userdata, envir = Envir)
if(as.character(list(Envir)) != "<environment>") stop ("No such environment")
if(exists("Envir", mode = "environment") == FALSE) stop ("No such environment")
FPCEnv$env <- Envir
testbounds <- 1
testpar <- 1
is.na(testbounds) <- TRUE
is.na(testpar) <- TRUE
testbounds <- try(get(pnoptnm, envir = Envir)[16:32],
silent = TRUE)
testpar <- try(get(pnoptnm, envir = Envir)[1:15],
silent = TRUE)
if (class(testbounds)[1] == "try-error" | class(testpar)[1] ==
"try-error" | is.na(testbounds[1]) == TRUE | is.na(testpar[1]) ==
TRUE)
try({
FPCEnv$mod.sel <- TRUE
modpar(datamerg[,1], datamerg[,2], pn.options = pnoptnm, taper.ends = taper.ends,
verbose=FALSE, Envir = Envir, ...)
options(warn=-1)
try(rm("mod.sel", envir = FPCEnv), silent =T)
options(warn=0)
}, silent = FALSE)
extraF <- try(get("extraF", pos = 1), silent = TRUE)
if (class(extraF)[1] == "try-error") {
stop("cannot find function: extraF - please reload FlexParamCurve")
}
print("checking fit of positive section of the curve for variable M*************************************")
richardsR12.lis <- try(get(richardsR12.lis, envir = FPCEnv),
silent = TRUE)
if (class(richardsR12.lis)[1] == "try-error" | existing ==
FALSE | (12 %in% mod.subset) == TRUE)
if(alacarte == FALSE | (12 %in% mod.subset) == TRUE){
richardsR12.lis <- eval(parse(text=sprintf("%s",paste("try(nlsList(y ~ SSposnegRichards(x,
Asym = Asym, K = K, Infl = Infl, M = M, modno = 12, pn.options = ",pnoptnm, "), data = userdata),
silent = TRUE)",sep=""))))}
print("checking fit of positive section of the curve for fixed M*************************************")
pnmodelparams <- get(pnoptnm, envir = Envir)[1:15]
change.pnparameters <- try(get("change.pnparameters", pos = 1),
silent = TRUE)
chk <- try({
todump<-unlist(summary(richardsR12.lis))["RSE"]
}, silent = TRUE)
richardsR20.lis <- try(get(richardsR20.lis, envir = FPCEnv),
silent = TRUE)
if (class(richardsR20.lis)[1] == "try-error" | existing ==
FALSE | (20 %in% mod.subset) == TRUE)
if(alacarte == FALSE | (20 %in% mod.subset) == TRUE){
richardsR20.lis <- eval(parse(text=sprintf("%s",paste("try(nlsList(y ~ SSposnegRichards(x,
Asym = Asym, K = K, Infl = Infl, modno = 20, pn.options = ",pnoptnm,"), data = userdata),
silent = TRUE)",sep=""))))}
chk1 <- try({
todump<-unlist(summary(richardsR20.lis))["RSE"]
}, silent = TRUE)
if ((class(richardsR20.lis)[1]) == "try-error"| class(chk1)[1] == "try-error"
| class(richardsR20.lis)[[1]] != "nlsList" ) {
print("3 parameter positive richards model failed/not fitted*************************************")
if(forcemod != 3) forcemod = 4
richardsR20.lis <- 1
} else {
FPCEnv$richardsR20.lis <- richardsR20.lis
}
if ((class(richardsR12.lis)[1]) == "try-error" | class(chk)[1] == "try-error"
| class(richardsR12.lis)[[1]] != "nlsList" )
{
print("4 parameter positive richards model failed/not fitted*************************************")
if(forcemod != 4) forcemod = 3
richardsR12.lis <- 1
} else {
FPCEnv$richardsR12.lis <- richardsR12.lis
}
currentmodel <- 1
testmod <- try(extraF(richardsR20.lis, richardsR12.lis, warn = F), silent = TRUE)
if (forcemod == 0) {
if (class(testmod)[1] == "try-error") {
modelsig = 0.1
} else {
modelsig = testmod[4]
if ((testmod[4]) > 0.05 & sqrt(testmod[5]/(testmod[3]-testmod[2])) > sqrt(testmod[6]/testmod[3])) {
currentmodel <- richardsR20.lis
mostreducednm <- substr("richardsR20.lis", 10,
11)
} else {
currentmodel <- richardsR12.lis
mostreducednm <- substr("richardsR12.lis", 10,
11)
}
}
}
mostreducedmod <- currentmodel
if (class(testmod)[1] != "try-error") {
mostreducednm <- substr("richardsR20.lis", 10,
11)
mostreducedmod <- richardsR20.lis
} else {
mostreducednm <- "NONE"
}
if (forcemod == 3)
{
modelsig = 0.1
}
if (forcemod == 4)
{
modelsig = 0.04
}
if (modelsig < 0.05) {
print("Variable M models most appropriate*************************************")
modno <- c(1:16)
} else {
print("Fixed M models most appropriate*************************************")
modno <- c(21:36)
}
if(alacarte == TRUE) modno = mod.subset
runmod <- function(userdata, modelno, modsig, existing = FALSE) {
savnm <- paste("richardsR", as.character(modelno), ".lis",
sep = "")
if (modelno[1] <20) {
savM <- ",M = M"
} else {
savM <- " "
}
if (modelno[1] < 17 | modelno[1] >= 18) {
savK <- ",K = K"
} else {
savK <- " "
}
mod <- 1
options(warn = -1)
modnmtry <- paste('FPCEnv$richardsR', modelno,'.lis', sep = "")
modexist <- try( eval(parse(text=sprintf("%s", modnmtry))), silent = TRUE)
options(warn = 0)
if (class(modexist)[1] == "try-error" | existing == FALSE) {
if ((modelno %in% mod.subset) == TRUE | alacarte == FALSE) {
if (modelno == 1 | modelno == 21)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",RAsym=RAsym,Rk=Rk,Ri=Ri,RM=RM,modno=",
modelno, ", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))),
silent = TRUE)
if (modelno == 2 | modelno == 22)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",RAsym=RAsym,Rk=Rk,Ri=Ri,modno=",
modelno, ", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))),
silent = TRUE)
if (modelno == 3 | modelno == 23 | modelno == 17.1)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",Ri=Ri,RM=RM,modno=", modelno,
", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))), silent = TRUE)
if (modelno == 4 | modelno == 24)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",RAsym=RAsym,RM=RM,modno=",
modelno, ", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))),
silent = TRUE)
if (modelno == 5 | modelno == 25)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",RM=RM,modno=", modelno,
", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))), silent = TRUE)
if (modelno == 6 | modelno == 26 | modelno == 17)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",RAsym=RAsym,Ri=Ri,RM=RM,modno=",
modelno, ", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))),
silent = TRUE)
if (modelno == 7 | modelno == 27)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",Rk=Rk,Ri=Ri,RM=RM,modno=",
modelno, ", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))),
silent = TRUE)
if (modelno == 8 | modelno == 28)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",RAsym=RAsym,Rk=Rk,RM=RM,modno=",
modelno, ", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))),
silent = TRUE)
if (modelno == 9 | modelno == 29)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",Rk=Rk,RM=RM,modno=", modelno,
", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))), silent = TRUE)
if (modelno == 10 | modelno == 30 | modelno == 17.3)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",Ri=Ri,modno=", modelno,
", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))), silent = TRUE)
if (modelno == 11 | modelno == 31)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",RAsym=RAsym,modno=",
modelno, ", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))),
silent = TRUE)
if (modelno == 12 | modelno == 32)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",modno=", modelno,
", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))), silent = TRUE)
if (modelno == 13 | modelno == 33 | modelno == 17.2)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",RAsym=RAsym,Ri=Ri,modno=",
modelno, ", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))),
silent = TRUE)
if (modelno == 14 | modelno == 34)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",Rk=Rk,Ri=Ri,modno=", modelno,
", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))), silent = TRUE)
if (modelno == 15 | modelno == 35)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",RAsym=RAsym,Rk=Rk,modno=",
modelno, ", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))),
silent = TRUE)
if (modelno == 16 | modelno == 36)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",Rk=Rk,modno=", modelno,
", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))), silent = TRUE)
}
} else {
options(warn = -1)
print("Model already exists in working environment")
mod <- try(eval(parse(text=sprintf("%s",paste("FPCEnv$",savnm,sep="")))), silent = TRUE)
options(warn = 0)
}
if (class(mod)[[1]] != "nlsList")
mod <- NULL
checkmod <- try(if (is.null(nrow(coef(mod))) == TRUE) {
mod <- NULL
}, silent = TRUE)
if (class(checkmod)[1] == "try-error" | class(mod)[1] ==
"NULL") {
messagesav <- (paste("********************** Model ",
savnm, " has not been successfully fit, please trouble-shoot this model separately and then repeat function using existing=TRUE *************************************************",
sep = ""))
} else {
if(existing != FALSE & class(modexist)[1] != "try-error") {
messagesav <- (paste("********************** Model ",
savnm, " exists and transferred successfully to the FlexParamCurve:::FPCEnv environment *************************************************",
sep = ""))
} else {
messagesav <- (paste("********************** Model ",
savnm, " fit successfully and saved in FlexParamCurve:::FPCEnv environment *************************************************",
sep = ""))
}
assign(savnm, mod, envir = as.environment(FPCEnv))
return(messagesav)
}
}
skel <- rep(list(1), 16)
initval <- c(rep(NA, 16))
initval <- relist(initval, skel)
for (i in 1:length(modno)) {
print("################ ################ ################## ################# ############### #########")
print(paste("Fitting model ",i," of ",length(modno),": richardsR",modno[i],".lis",sep=""))
initval[i] <- runmod(userdata, modno[i], modelsig, existing = existing)
print(initval[i])
}
print("#########################################################################################################")
print("Model fitting completed, summary of fits follows:")
print(initval[1:length(modno)])
print("#########################################################################################################")
print("Writing summary tables.....")
modrank <- data.frame(`PN Richards Model` = rep(NA, length(modno)),
`Ranking function value` = rep(-999, length(modno)),
`No. Individuals Fitted` = rep(-999, length(modno)),
RSE = rep(-999, length(modno)), `df model` = rep(-999,
length(modno)), `df residual` = rep(-999, length(modno)),
`No. curve params` = rep(-999, length(modno)))
for (i in 1:length(modno)) {
modnm <- paste("richardsR", as.character(modno[i]), ".lis",
sep = "")
RSEstr <- "RSE"
dfstr <- "df"
dfstr1 <- "df.residual"
usefun <- unlist(strsplit(penaliz, "(n)"))
FPCEnv$model1 <- try(get(modnm, envir = as.environment(FPCEnv)), silent = TRUE)
if(class(FPCEnv$model1)[1] == "try-error") {
FPCEnv$model1 <- NULL
checkmod <- NULL
} else {
checkmod <- try(if (is.null(nrow(coef(FPCEnv$model1))) == TRUE) {
FPCEnv$model1 <- NULL
} else {
FPCEnv$model1 <- FPCEnv$model1
}, silent = TRUE)
}
evfun <- parse(text = sprintf("%s", paste("summary(FPCEnv$model1)[['",
RSEstr, "']]*(", usefun[1], "(1+sum( summary(FPCEnv$model1)[['",
dfstr, "']],na.rm=TRUE)))", usefun[2], sep = "")))
if (class(FPCEnv$model1)[1] == "nlsList" & class(checkmod)[1] !=
"try-error") {
modrank[i, 2] <- eval(evfun)
modrank[i, 3] <- nrow(subset(coef(FPCEnv$model1)[1], is.na((coef(FPCEnv$model1)[1])) ==
FALSE))
modrank[i, 4] <- eval(parse(text = sprintf("%s",
paste("summary(FPCEnv$model1)[['", RSEstr, "']]", sep = ""))))
modrank[i, 5] <- eval(parse(text = sprintf("%s",
paste("sum(summary(FPCEnv$model1)[['", dfstr, "']][,1],na.rm=TRUE)",
sep = ""))))
modrank[i, 6] <- eval(parse(text = sprintf("%s",
paste("sum(summary(FPCEnv$model1)[['", dfstr1, "']],na.rm=TRUE)",
sep = ""))))
modrank[i, 7] <- length(coef(FPCEnv$model1))
} else {
}
modrank[i, 1] <- modnm
}
pvalmat <- data.frame(matrix(nrow = length(modno), ncol = length(modno)))
if(length(modno) >1) {
for (i in 1:length(modno)) {
for (j in 2:(length(modno) - 1)) {
if (modrank[i, 2] != -999 & modrank[j, 2] != -999) {
mod1 <- eval(parse(text = sprintf("%s", paste("FPCEnv$", modrank[i,
1], sep = ""))))
mod2 <- eval(parse(text = sprintf("%s", paste("FPCEnv$", modrank[j,
1], sep = ""))))
if (length(coef(mod1)) > length(coef(mod2))) {
nmmod1 <- names(coef(mod1))
nmmod2 <- names(coef(mod2))
if (length(nmmod2 %in% nmmod1) == length(nmmod2))
pvalmat[i, j] <- as.numeric(extraF(mod2,
mod1, warn = F)[4])
if (length(nmmod2 %in% nmmod1) == length(nmmod2))
pvalmat[j, i] <- pvalmat[i, j]
} else {
if (length(coef(mod1)) < length(coef(mod2))) {
nmmod1 <- names(coef(mod1))
nmmod2 <- names(coef(mod2))
if (length(nmmod1 %in% nmmod2) == length(nmmod1))
pvalmat[i, j] <- as.numeric(extraF(mod1,
mod2, warn = F)[4])
if (length(nmmod1 %in% nmmod2) == length(nmmod1))
pvalmat[j, i] <- pvalmat[i, j]
} else {
}
}
} else {
}
}
if (modrank[i, 2] != -999) {
modtemp <- eval(parse(text = sprintf("%s", paste("FPCEnv$", modrank[i,
1], sep =""))))
modnm <- paste("R", as.character(modno[i]), ".lis", sprintf("(%s)",
length(coef(modtemp))), sep = "")
row.names(pvalmat)[i] <- paste("R", as.character(modno[i]),
sprintf("(%s)", length(coef(modtemp))), sep = "")
} else {
modtemp <- NULL
modnm <- paste("R", as.character(modno[i]), ".lis", sprintf("(%s)",0), sep = "")
row.names(pvalmat)[i] <- paste("R", as.character(modno[i]), ".lis",
sprintf("(%s)",0), sep = "")
}
names(pvalmat)[i] <- modnm
}
}
pvalmat <- apply(pvalmat, 1, function(x) round(x, 3))
modrank[modrank == -999] <- NA
modrank[is.na(modrank[,2]),2] <- Inf
modrank[is.na(modrank[,6]),6] <- Inf
modrank <- modrank[order(modrank[, 2], modrank[, 6]), ]
modrank[(modrank[,2])==Inf,2] <- NA
modrank[(modrank[,6])==Inf,6] <- NA
row.names(modrank) <- c(1:length(modno))
print("...done")
print("writing output to environment:")
print(Envir)
get.mod(to.envir = Envir, write.mod = TRUE)
assign("userdata",userdata, envir = Envir)
outp <- list(modrank, pvalmat)
names(outp) <- c("Model rank table", "P values from pairwise extraF comparisons")
options(warn=-1)
try(rm("model1",envir = FPCEnv),silent=T)
try(rm("tempparam.select",envir = FPCEnv),silent=T)
try(rm("tempmodnm",envir = FPCEnv),silent=T)
try(rm("legitmodel",envir = FPCEnv),silent=T)
options(warn = 0)
return(outp)
##value<< A list object with two components: $'Model rank table' contains the
## statistics from \code{\link{extraF}} ranked by the modified residual standard error,
## and $'P values from pairwise extraF comparison' is a matrix of P values from
## \code{\link{extraF}} for legitimate comparisons (nested and successfully fitted models).
## The naming convention for models is a concatenation of 'richardsR', the modno and '.lis'
## which is shortened in the matrix output, where the number of parameters has been
## pasted in parentheses to allow users to easily distinguish the more general model from
## the more reduced model
## (see \code{\link{extraF}} and \code{\link{SSposnegRichards}}).
##seealso<< \code{\link{extraF}}
## \code{\link{SSposnegRichards}}
## \code{\link{nlsList}}
##note<< If object \eqn{pnmodelparams} does not exist, \code{\link{modpar}}
## will be called automatically prior to model selection. During selection, text is output
## to the screen to inform the user of the progress of model selection
## (which model is being fitted, which were fitted successfully)
}
, ex = function(){
#run model selection for posneg.data object (only first 3 group levels for example's sake)
subdata <- subset(posneg.data, as.numeric(row.names (posneg.data) ) < 40)
modseltable <- pn.mod.compare(subdata$age, subdata$mass,
subdata$id, existing = FALSE, pn.options = "myoptions")
#fit nlsList model initially and then run model selection
#for posneg.data object when at least one model is already fit
# note forcemod is set to 3 so that models 21-36 are evaluated
subdata <- subset(posneg.data, as.numeric(row.names (posneg.data) ) < 40)
richardsR22.lis <- nlsList(mass ~ SSposnegRichards(age, Asym = Asym, K = K,
Infl = Infl, RAsym = RAsym, Rk = Rk, Ri = Ri, modno = 22)
,data = posneg.data)
modseltable <- pn.mod.compare(subdata$age, subdata$mass,
subdata$id, forcemod = 3, existing = TRUE)
#run model selection ranked by residual standard error*sample size
modseltable <- pn.mod.compare(subdata$age, subdata$mass,
subdata$id, penaliz='1*(n)', existing = TRUE)
}
)
Try the FlexParamCurve package in your browser
Any scripts or data that you put into this service are public.
FlexParamCurve documentation built on May 1, 2019, 11:36 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.
pn.mod.compare <-
structure(function # Compare All Possible Positive-Negative Richards \eqn{nlslist} Models
(x,
### a numeric vector of the primary predictor
y,
### a numeric vector of the response variable
grp,
### a factor of same length as x and y that distinguishes groups within
### the dataset
pn.options,
### required character value for name of list object populated with starting
### parameter estimates, fitting
### options and bounds or destination for modpar to write a new list (see Details)
forcemod = 0,
### optional numeric value to constrain model selection (see Details)
existing = FALSE,
### optional logical value specifying whether some of the relevant models
### have already been fitted
penaliz = "1/sqrt(n)",
### optional character value to determine how models are ranked (see Details)
taper.ends = 0.45,
mod.subset = c(NA),
Envir = .GlobalEnv,
...
) {
##description<< This function performs model selection for \code{\link{nlsList}} models fitted using
## \code{\link{SSposnegRichards}}.
##details<< First, whether parameter M should be fixed
## (see \code{\link{SSposnegRichards}}) is determined by fitting models 12 and 20 and comparing
## their perfomance using \code{\link{extraF}}.
## If model 12 provides superior performance (variable values of M) then 16 models that estimate M
## are run
## (models 1 through 16), otherwise the models with fixed M are fitted (models 21 through 36).
## Fitting these \code{\link{nlsList}} models can be time-consuming (2-4 hours using the dataset
## \code{\link{posneg.data}} that encompasses 100 individuals) and if several of the relevant
## models are already fitted the option existing=TRUE can be used to avoid refitting models that
## already exist (note that a model object in which no grouping levels were successfully
## parameterized will be refitted, as will objects that are not of class nlsList).
##
## Specifying forcemod=3 will force model selection to only consider fixed M models and setting
## forcemod=4 will force model selection to consider models with varying values of M only.
## If fitting both models
## 12 and 20 fails, fixed M models will be used by default.
##
## Models are ranked by modified pooled residual square error. By default residual standard error
## is divided by the square root of sample size. This exponentially penalizes models for which very few
## grouping levels (individuals) are successfully parameterized (the few individuals that are
## parameterized in these models are fit unsuprisingly well) using a function based on the relationship
## between standard error and sample size. However, different users may have different preferences
## and these can be specified in the argument penaliz (which residual
## standard error is multiplied by). This argument must be a character value
## that contains the character n (sample size) and must be a valid right hand side (RHS) of a formula:
## e.g. 1*(n), (n)^2. It cannot contain more than one n but could be a custom function, e.g. FUN(n).
if( anyNA(data.frame(x, y, grp)) == TRUE ) stop ("This function does not handle missing data values for x, y, or grp. Please subset the data, e.g. mydataframe[!is.na(mydataframe),], to remove them prior to function call")
options(warn = -1)
if(existing == FALSE) {
rm(list = ls(pattern = "richardsR", envir=FPCEnv), envir=FPCEnv)
rm(list = ls(pattern = "richardsR", envir=Envir), envir=Envir)
} else {
rm(list = ls(pattern = "richardsR", envir=FPCEnv), envir=FPCEnv)
get.mod(modelname = ls(Envir,pattern="richardsR"), from.envir = Envir, to.envir = FPCEnv, write.mod = TRUE, silent = TRUE)
print("#########################################################################################################")
print(paste("NOTE: existing is not set to false, existing models in the working environment ",substitute(Envir)," will be used during fits. To remove models manually, remove all files prefixed richardsR from working environment before running",sep=""))
print("#########################################################################################################")
}
alacarte <- FALSE
"%w/o%" <- function(x, y) x[!x %in% y] #-- x without y
if(length(mod.subset %w/o% NA) > 0) {
alacarte <- TRUE
}
pnoptnm <- as.character(pn.options[1])
checkpen <- try(unlist(strsplit(penaliz, "(n)")), silent = TRUE)
if (length(checkpen) != 2 | class(checkpen)[1] == "try-error") {
stop("penaliz parameter is ill defined: see ?pn.mod.compare")
} else {
checkpen <- try(eval(parse(text = sprintf("%s", paste(checkpen[1],
"1", checkpen[2], sep = "")))))
if (class(checkpen)[1] == "try-error")
stop("penaliz parameter is ill defined: see ?pn.mod.compare")
}
datamerg <- data.frame(x, y, grp)
userdata <- groupedData(y ~ x | grp, outer = ~grp, data = datamerg)
assign("userdata",userdata, envir = Envir)
if(as.character(list(Envir)) != "<environment>") stop ("No such environment")
if(exists("Envir", mode = "environment") == FALSE) stop ("No such environment")
FPCEnv$env <- Envir
testbounds <- 1
testpar <- 1
is.na(testbounds) <- TRUE
is.na(testpar) <- TRUE
testbounds <- try(get(pnoptnm, envir = Envir)[16:32],
silent = TRUE)
testpar <- try(get(pnoptnm, envir = Envir)[1:15],
silent = TRUE)
if (class(testbounds)[1] == "try-error" | class(testpar)[1] ==
"try-error" | is.na(testbounds[1]) == TRUE | is.na(testpar[1]) ==
TRUE)
try({
FPCEnv$mod.sel <- TRUE
modpar(datamerg[,1], datamerg[,2], pn.options = pnoptnm, taper.ends = taper.ends,
verbose=FALSE, Envir = Envir, ...)
options(warn=-1)
try(rm("mod.sel", envir = FPCEnv), silent =T)
options(warn=0)
}, silent = FALSE)
extraF <- try(get("extraF", pos = 1), silent = TRUE)
if (class(extraF)[1] == "try-error") {
stop("cannot find function: extraF - please reload FlexParamCurve")
}
print("checking fit of positive section of the curve for variable M*************************************")
richardsR12.lis <- try(get(richardsR12.lis, envir = FPCEnv),
silent = TRUE)
if (class(richardsR12.lis)[1] == "try-error" | existing ==
FALSE | (12 %in% mod.subset) == TRUE)
if(alacarte == FALSE | (12 %in% mod.subset) == TRUE){
richardsR12.lis <- eval(parse(text=sprintf("%s",paste("try(nlsList(y ~ SSposnegRichards(x,
Asym = Asym, K = K, Infl = Infl, M = M, modno = 12, pn.options = ",pnoptnm, "), data = userdata),
silent = TRUE)",sep=""))))}
print("checking fit of positive section of the curve for fixed M*************************************")
pnmodelparams <- get(pnoptnm, envir = Envir)[1:15]
change.pnparameters <- try(get("change.pnparameters", pos = 1),
silent = TRUE)
chk <- try({
todump<-unlist(summary(richardsR12.lis))["RSE"]
}, silent = TRUE)
richardsR20.lis <- try(get(richardsR20.lis, envir = FPCEnv),
silent = TRUE)
if (class(richardsR20.lis)[1] == "try-error" | existing ==
FALSE | (20 %in% mod.subset) == TRUE)
if(alacarte == FALSE | (20 %in% mod.subset) == TRUE){
richardsR20.lis <- eval(parse(text=sprintf("%s",paste("try(nlsList(y ~ SSposnegRichards(x,
Asym = Asym, K = K, Infl = Infl, modno = 20, pn.options = ",pnoptnm,"), data = userdata),
silent = TRUE)",sep=""))))}
chk1 <- try({
todump<-unlist(summary(richardsR20.lis))["RSE"]
}, silent = TRUE)
if ((class(richardsR20.lis)[1]) == "try-error"| class(chk1)[1] == "try-error"
| class(richardsR20.lis)[[1]] != "nlsList" ) {
print("3 parameter positive richards model failed/not fitted*************************************")
if(forcemod != 3) forcemod = 4
richardsR20.lis <- 1
} else {
FPCEnv$richardsR20.lis <- richardsR20.lis
}
if ((class(richardsR12.lis)[1]) == "try-error" | class(chk)[1] == "try-error"
| class(richardsR12.lis)[[1]] != "nlsList" )
{
print("4 parameter positive richards model failed/not fitted*************************************")
if(forcemod != 4) forcemod = 3
richardsR12.lis <- 1
} else {
FPCEnv$richardsR12.lis <- richardsR12.lis
}
currentmodel <- 1
testmod <- try(extraF(richardsR20.lis, richardsR12.lis, warn = F), silent = TRUE)
if (forcemod == 0) {
if (class(testmod)[1] == "try-error") {
modelsig = 0.1
} else {
modelsig = testmod[4]
if ((testmod[4]) > 0.05 & sqrt(testmod[5]/(testmod[3]-testmod[2])) > sqrt(testmod[6]/testmod[3])) {
currentmodel <- richardsR20.lis
mostreducednm <- substr("richardsR20.lis", 10,
11)
} else {
currentmodel <- richardsR12.lis
mostreducednm <- substr("richardsR12.lis", 10,
11)
}
}
}
mostreducedmod <- currentmodel
if (class(testmod)[1] != "try-error") {
mostreducednm <- substr("richardsR20.lis", 10,
11)
mostreducedmod <- richardsR20.lis
} else {
mostreducednm <- "NONE"
}
if (forcemod == 3)
{
modelsig = 0.1
}
if (forcemod == 4)
{
modelsig = 0.04
}
if (modelsig < 0.05) {
print("Variable M models most appropriate*************************************")
modno <- c(1:16)
} else {
print("Fixed M models most appropriate*************************************")
modno <- c(21:36)
}
if(alacarte == TRUE) modno = mod.subset
runmod <- function(userdata, modelno, modsig, existing = FALSE) {
savnm <- paste("richardsR", as.character(modelno), ".lis",
sep = "")
if (modelno[1] <20) {
savM <- ",M = M"
} else {
savM <- " "
}
if (modelno[1] < 17 | modelno[1] >= 18) {
savK <- ",K = K"
} else {
savK <- " "
}
mod <- 1
options(warn = -1)
modnmtry <- paste('FPCEnv$richardsR', modelno,'.lis', sep = "")
modexist <- try( eval(parse(text=sprintf("%s", modnmtry))), silent = TRUE)
options(warn = 0)
if (class(modexist)[1] == "try-error" | existing == FALSE) {
if ((modelno %in% mod.subset) == TRUE | alacarte == FALSE) {
if (modelno == 1 | modelno == 21)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",RAsym=RAsym,Rk=Rk,Ri=Ri,RM=RM,modno=",
modelno, ", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))),
silent = TRUE)
if (modelno == 2 | modelno == 22)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",RAsym=RAsym,Rk=Rk,Ri=Ri,modno=",
modelno, ", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))),
silent = TRUE)
if (modelno == 3 | modelno == 23 | modelno == 17.1)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",Ri=Ri,RM=RM,modno=", modelno,
", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))), silent = TRUE)
if (modelno == 4 | modelno == 24)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",RAsym=RAsym,RM=RM,modno=",
modelno, ", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))),
silent = TRUE)
if (modelno == 5 | modelno == 25)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",RM=RM,modno=", modelno,
", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))), silent = TRUE)
if (modelno == 6 | modelno == 26 | modelno == 17)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",RAsym=RAsym,Ri=Ri,RM=RM,modno=",
modelno, ", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))),
silent = TRUE)
if (modelno == 7 | modelno == 27)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",Rk=Rk,Ri=Ri,RM=RM,modno=",
modelno, ", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))),
silent = TRUE)
if (modelno == 8 | modelno == 28)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",RAsym=RAsym,Rk=Rk,RM=RM,modno=",
modelno, ", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))),
silent = TRUE)
if (modelno == 9 | modelno == 29)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",Rk=Rk,RM=RM,modno=", modelno,
", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))), silent = TRUE)
if (modelno == 10 | modelno == 30 | modelno == 17.3)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",Ri=Ri,modno=", modelno,
", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))), silent = TRUE)
if (modelno == 11 | modelno == 31)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",RAsym=RAsym,modno=",
modelno, ", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))),
silent = TRUE)
if (modelno == 12 | modelno == 32)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",modno=", modelno,
", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))), silent = TRUE)
if (modelno == 13 | modelno == 33 | modelno == 17.2)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",RAsym=RAsym,Ri=Ri,modno=",
modelno, ", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))),
silent = TRUE)
if (modelno == 14 | modelno == 34)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",Rk=Rk,Ri=Ri,modno=", modelno,
", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))), silent = TRUE)
if (modelno == 15 | modelno == 35)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",RAsym=RAsym,Rk=Rk,modno=",
modelno, ", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))),
silent = TRUE)
if (modelno == 16 | modelno == 36)
mod <- try(eval(parse(text = sprintf("%s", paste("nlsList(y~SSposnegRichards(x,Asym=Asym",savK,",Infl=Infl",
savM, ",Rk=Rk,modno=", modelno,
", pn.options = \"",pnoptnm, "\"),data=userdata, ...)", sep = "")))), silent = TRUE)
}
} else {
options(warn = -1)
print("Model already exists in working environment")
mod <- try(eval(parse(text=sprintf("%s",paste("FPCEnv$",savnm,sep="")))), silent = TRUE)
options(warn = 0)
}
if (class(mod)[[1]] != "nlsList")
mod <- NULL
checkmod <- try(if (is.null(nrow(coef(mod))) == TRUE) {
mod <- NULL
}, silent = TRUE)
if (class(checkmod)[1] == "try-error" | class(mod)[1] ==
"NULL") {
messagesav <- (paste("********************** Model ",
savnm, " has not been successfully fit, please trouble-shoot this model separately and then repeat function using existing=TRUE *************************************************",
sep = ""))
} else {
if(existing != FALSE & class(modexist)[1] != "try-error") {
messagesav <- (paste("********************** Model ",
savnm, " exists and transferred successfully to the FlexParamCurve:::FPCEnv environment *************************************************",
sep = ""))
} else {
messagesav <- (paste("********************** Model ",
savnm, " fit successfully and saved in FlexParamCurve:::FPCEnv environment *************************************************",
sep = ""))
}
assign(savnm, mod, envir = as.environment(FPCEnv))
return(messagesav)
}
}
skel <- rep(list(1), 16)
initval <- c(rep(NA, 16))
initval <- relist(initval, skel)
for (i in 1:length(modno)) {
print("################ ################ ################## ################# ############### #########")
print(paste("Fitting model ",i," of ",length(modno),": richardsR",modno[i],".lis",sep=""))
initval[i] <- runmod(userdata, modno[i], modelsig, existing = existing)
print(initval[i])
}
print("#########################################################################################################")
print("Model fitting completed, summary of fits follows:")
print(initval[1:length(modno)])
print("#########################################################################################################")
print("Writing summary tables.....")
modrank <- data.frame(`PN Richards Model` = rep(NA, length(modno)),
`Ranking function value` = rep(-999, length(modno)),
`No. Individuals Fitted` = rep(-999, length(modno)),
RSE = rep(-999, length(modno)), `df model` = rep(-999,
length(modno)), `df residual` = rep(-999, length(modno)),
`No. curve params` = rep(-999, length(modno)))
for (i in 1:length(modno)) {
modnm <- paste("richardsR", as.character(modno[i]), ".lis",
sep = "")
RSEstr <- "RSE"
dfstr <- "df"
dfstr1 <- "df.residual"
usefun <- unlist(strsplit(penaliz, "(n)"))
FPCEnv$model1 <- try(get(modnm, envir = as.environment(FPCEnv)), silent = TRUE)
if(class(FPCEnv$model1)[1] == "try-error") {
FPCEnv$model1 <- NULL
checkmod <- NULL
} else {
checkmod <- try(if (is.null(nrow(coef(FPCEnv$model1))) == TRUE) {
FPCEnv$model1 <- NULL
} else {
FPCEnv$model1 <- FPCEnv$model1
}, silent = TRUE)
}
evfun <- parse(text = sprintf("%s", paste("summary(FPCEnv$model1)[['",
RSEstr, "']]*(", usefun[1], "(1+sum( summary(FPCEnv$model1)[['",
dfstr, "']],na.rm=TRUE)))", usefun[2], sep = "")))
if (class(FPCEnv$model1)[1] == "nlsList" & class(checkmod)[1] !=
"try-error") {
modrank[i, 2] <- eval(evfun)
modrank[i, 3] <- nrow(subset(coef(FPCEnv$model1)[1], is.na((coef(FPCEnv$model1)[1])) ==
FALSE))
modrank[i, 4] <- eval(parse(text = sprintf("%s",
paste("summary(FPCEnv$model1)[['", RSEstr, "']]", sep = ""))))
modrank[i, 5] <- eval(parse(text = sprintf("%s",
paste("sum(summary(FPCEnv$model1)[['", dfstr, "']][,1],na.rm=TRUE)",
sep = ""))))
modrank[i, 6] <- eval(parse(text = sprintf("%s",
paste("sum(summary(FPCEnv$model1)[['", dfstr1, "']],na.rm=TRUE)",
sep = ""))))
modrank[i, 7] <- length(coef(FPCEnv$model1))
} else {
}
modrank[i, 1] <- modnm
}
pvalmat <- data.frame(matrix(nrow = length(modno), ncol = length(modno)))
if(length(modno) >1) {
for (i in 1:length(modno)) {
for (j in 2:(length(modno) - 1)) {
if (modrank[i, 2] != -999 & modrank[j, 2] != -999) {
mod1 <- eval(parse(text = sprintf("%s", paste("FPCEnv$", modrank[i,
1], sep = ""))))
mod2 <- eval(parse(text = sprintf("%s", paste("FPCEnv$", modrank[j,
1], sep = ""))))
if (length(coef(mod1)) > length(coef(mod2))) {
nmmod1 <- names(coef(mod1))
nmmod2 <- names(coef(mod2))
if (length(nmmod2 %in% nmmod1) == length(nmmod2))
pvalmat[i, j] <- as.numeric(extraF(mod2,
mod1, warn = F)[4])
if (length(nmmod2 %in% nmmod1) == length(nmmod2))
pvalmat[j, i] <- pvalmat[i, j]
} else {
if (length(coef(mod1)) < length(coef(mod2))) {
nmmod1 <- names(coef(mod1))
nmmod2 <- names(coef(mod2))
if (length(nmmod1 %in% nmmod2) == length(nmmod1))
pvalmat[i, j] <- as.numeric(extraF(mod1,
mod2, warn = F)[4])
if (length(nmmod1 %in% nmmod2) == length(nmmod1))
pvalmat[j, i] <- pvalmat[i, j]
} else {
}
}
} else {
}
}
if (modrank[i, 2] != -999) {
modtemp <- eval(parse(text = sprintf("%s", paste("FPCEnv$", modrank[i,
1], sep =""))))
modnm <- paste("R", as.character(modno[i]), ".lis", sprintf("(%s)",
length(coef(modtemp))), sep = "")
row.names(pvalmat)[i] <- paste("R", as.character(modno[i]),
sprintf("(%s)", length(coef(modtemp))), sep = "")
} else {
modtemp <- NULL
modnm <- paste("R", as.character(modno[i]), ".lis", sprintf("(%s)",0), sep = "")
row.names(pvalmat)[i] <- paste("R", as.character(modno[i]), ".lis",
sprintf("(%s)",0), sep = "")
}
names(pvalmat)[i] <- modnm
}
}
pvalmat <- apply(pvalmat, 1, function(x) round(x, 3))
modrank[modrank == -999] <- NA
modrank[is.na(modrank[,2]),2] <- Inf
modrank[is.na(modrank[,6]),6] <- Inf
modrank <- modrank[order(modrank[, 2], modrank[, 6]), ]
modrank[(modrank[,2])==Inf,2] <- NA
modrank[(modrank[,6])==Inf,6] <- NA
row.names(modrank) <- c(1:length(modno))
print("...done")
print("writing output to environment:")
print(Envir)
get.mod(to.envir = Envir, write.mod = TRUE)
assign("userdata",userdata, envir = Envir)
outp <- list(modrank, pvalmat)
names(outp) <- c("Model rank table", "P values from pairwise extraF comparisons")
options(warn=-1)
try(rm("model1",envir = FPCEnv),silent=T)
try(rm("tempparam.select",envir = FPCEnv),silent=T)
try(rm("tempmodnm",envir = FPCEnv),silent=T)
try(rm("legitmodel",envir = FPCEnv),silent=T)
options(warn = 0)
return(outp)
##value<< A list object with two components: $'Model rank table' contains the
## statistics from \code{\link{extraF}} ranked by the modified residual standard error,
## and $'P values from pairwise extraF comparison' is a matrix of P values from
## \code{\link{extraF}} for legitimate comparisons (nested and successfully fitted models).
## The naming convention for models is a concatenation of 'richardsR', the modno and '.lis'
## which is shortened in the matrix output, where the number of parameters has been
## pasted in parentheses to allow users to easily distinguish the more general model from
## the more reduced model
## (see \code{\link{extraF}} and \code{\link{SSposnegRichards}}).
##seealso<< \code{\link{extraF}}
## \code{\link{SSposnegRichards}}
## \code{\link{nlsList}}
##note<< If object \eqn{pnmodelparams} does not exist, \code{\link{modpar}}
## will be called automatically prior to model selection. During selection, text is output
## to the screen to inform the user of the progress of model selection
## (which model is being fitted, which were fitted successfully)
}
, ex = function(){
#run model selection for posneg.data object (only first 3 group levels for example's sake)
subdata <- subset(posneg.data, as.numeric(row.names (posneg.data) ) < 40)
modseltable <- pn.mod.compare(subdata$age, subdata$mass,
subdata$id, existing = FALSE, pn.options = "myoptions")
#fit nlsList model initially and then run model selection
#for posneg.data object when at least one model is already fit
# note forcemod is set to 3 so that models 21-36 are evaluated
subdata <- subset(posneg.data, as.numeric(row.names (posneg.data) ) < 40)
richardsR22.lis <- nlsList(mass ~ SSposnegRichards(age, Asym = Asym, K = K,
Infl = Infl, RAsym = RAsym, Rk = Rk, Ri = Ri, modno = 22)
,data = posneg.data)
modseltable <- pn.mod.compare(subdata$age, subdata$mass,
subdata$id, forcemod = 3, existing = TRUE)
#run model selection ranked by residual standard error*sample size
modseltable <- pn.mod.compare(subdata$age, subdata$mass,
subdata$id, penaliz='1*(n)', existing = TRUE)
}
)
Try the FlexParamCurve 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.