How to do fitting for each subsets of data and plot their fitting curves?

Question

Actually I have done much of the thing before I ask here and done lots of search but couldn't find the solution so far.

I have a subset of data that I need to fit each subset. On the other hand, when I try to plot regression lines of each of these subsets only one regression line is shown on the plot. That is the main problem so far.

Anyway lets go step by step; here is a data.frame

   xx <- rep(rep(seq(0,800,200),each=10),times=2)
   yy<-c(replicate(2,sort(10^runif(10,-1,0),decreasing=TRUE)),replicate(2,sort(10^runif(10,-1,0),decreasing=TRUE)), replicate(2,sort(10^runif(10,-2,0),decreasing=TRUE)),replicate(2,sort(10^runif(10,-3,0),decreasing=TRUE)), replicate(2,sort(10^runif(10,-4,0), decreasing=TRUE)))    
   V <- rep(seq(100,2500,length.out=10),times=2)
   No <- rep(1:10,each=10)
   df <- data.frame(V,xx,yy,No)

from wide format to long format

library(reshape2)
df_new <- melt(df,id=c("No","xx","V"))

**fitting model

As <- 220
Ax <- 1500
 
  model <- function(data){
  nlsLM(value~ifelse(V<Vs, 1-exp(-D*(1-(xx-As)/Ax)^2*(1-V/Vs)^2),1),
        data=data, start=c(D=1.21,Vs=1951),trace=T,control = nls.lm.control(maxiter=50))
}



library(plyr)
library(minpack.lm)

fit<- dlply(df_new, "No", .fun = model)

function for predicting fitting values ##from Adding Fitted Lines from an Existing Model

predictvals <- function(model, xvar, yvar, xrange=NULL, samples=10, ...) {
  # If xrange isn't passed in, determine xrange from the models.
  # Different ways of extracting the x range, depending on model type
  if (is.null(xrange)) {
    if (any(class(model) %in% c("nls", "glm")))
      xrange <- range(model$model[[xvar]])
    else if (any(class(model) %in% "loess"))
      xrange <- range(model$x)
  }
  newdata <- data.frame(x = seq(xrange[1],xrange[2], length.out = samples))
  names(newdata) <- xvar
  newdata[[yvar]] <- predict(model, newdata = newdata, ...)
  newdata
}

to form prediction lines that have the same x range across all groups

predvals<- ldply(fit, .fun=predictvals, xvar="V", yvar="value",xrange=range(df_new$V))
predvals$xx <- rep(rep(unique(df_new$xx),each=1),each=10)

and finally plot with facet_wrap~xx

ggplot(df_new,aes(y=value,x=V, col=factor(xx)))+
  geom_point(size=3,alpha=.4)+
  geom_line(data=predvals,aes(col=factor(xx)))+
  scale_y_log10(breaks=c(1e-6,1e-5,1e-4,1e-3,1e-2,1e-1,1),limits = c(1e-6,1))+
  facet_wrap(~xx,scales="free_x")

why there is only one fitting line. I expect two fitting line since I have two subset of data for each xx value. Which part I'm missing? any guidance will be appreciated.

Answer 1

Try adding another grouping factor to the aes call in geom_line.

... +
geom_line(data=predvals,aes(col=factor(xx), group=interaction(xx, No))) +
...

I haven't looked through the details of your code, so let me know if this isn't what you were looking for.