R中散点图矩阵的标签

Question

I am new to R, so please consider this when investigating my presumably unelegant attempts to problem-solving.

The data which I use come from a behavioral psychological study. I have a data frame, containing 6 different parameters (called mu, sigma, tau, v, a, Ter) from 3 different behavioral tasks each. Additionally, there are 2 stimulus types, face and house pictures. colnames(plot_data) yields the following:

 [1] "subject"                      "v.DelayedNMatchingFaces"     
 [3] "a.DelayedNMatchingFaces"      "Ter.DelayedNMatchingFaces"   
 [5] "mu.DelayedNMatchingFaces"     "sigma.DelayedNMatchingFaces" 
 [7] "tau.DelayedNMatchingFaces"    "v.DelayedNMatchingHouses"    
 [9] "a.DelayedNMatchingHouses"     "Ter.DelayedNMatchingHouses"  
[11] "mu.DelayedNMatchingHouses"    "sigma.DelayedNMatchingHouses"
[13] "tau.DelayedNMatchingHouses"   "v.MorphFaces"                
[15] "a.MorphFaces"                 "Ter.MorphFaces"              
[17] "mu.MorphFaces"                "sigma.MorphFaces"            
[19] "tau.MorphFaces"               "v.MorphHouses"               
[21] "a.MorphHouses"                "Ter.MorphHouses"             
[23] "mu.MorphHouses"               "sigma.MorphHouses"           
[25] "tau.MorphHouses"              "v.verificFaces"              
[27] "a.verificFaces"               "Ter.verificFaces"            
[29] "mu.verificFaces"              "sigma.verificFaces"          
[31] "tau.verificFaces"             "v.verificHouses"             
[33] "a.verificHouses"              "Ter.verificHouses"           
[35] "mu.verificHouses"             "sigma.verificHouses"         
[37] "tau.verificHouses"

Now, I need to do bivaraite scatterplots that contain one single parameter, but of each task, so I want to do a 3x3 scatterplot matrix (eg: a.DelayedNMatchingFaces , a.MorphFaces , a.verificFaces ). This matrix will be saved to a png that is named according to the parameter and the stimulus type (so for example "a_H.png" if it's the parameter a for house-stimuli). I have built a for-loop that goes through the columns and builds scatterplot a matrix of the variable from column i , column i+12 and column i+24 , because each parameter appears every 12th step in the header row. The loop looks as follows:

for (col_ix in 2:13) {

  i <- col_ix + 12
  j <- col_ix + 24

  param_col <- colnames(plot_data)[col_ix]
  param <- strsplit(param_col,"[.]")[[1]][1] 

  if (grepl("Faces", colnames(plot_data)[col_ix]) == TRUE) {
    stim_col <- "F" 
    }
  else {
    stim_col <- "H" 
    }

  png(paste(param, stim_col, ".png", sep="_"))
  scatterplotMatrix(~ plot_data[ ,col_ix] + plot_data[ ,i] + plot_data[ ,j] , span=0.7, data= plot_data)
  dev.off()

}

This seems to work, although I acknowledge that there must be a thousand ways to do it more efficiently, but hey, it's my first R loop, actually. The problem is: When I save the scatterplot matrix as png file, the plots that are visible in the pngs are labeled according to the data source that is indicated in the formula, so in this case, it would be plot_data[col_ix] , plot_data[i] and plot_data[j] (see below). The plots are supposed, however, to be labeled according to the column names, so that when you look at them, you can see which variable is displayed and plotted with which. There is an argument that is called var.label s, I saw that already, but I just can't figure out how to have it name every plot, because the indices col_ix , i and j are running numbers and cannot all three be listed in the var.labels argument.

Can anyone help? Suggestions are greatly appreciated (also, by the way, if they contain advice not directly linked to the variable labels). I apologize if my explanations were unprofessional or lacked the correct use of specific terms; hopefully everything was understandable. If there is more information required, please tell me, then I will add it.

Much thanks to you all.

Answer 1

you could just use the names associated with ij and col_ix as follows:

plot_data[ , names(plot_data)[col_ix]] + plot_data[ ,names(plot_data)[i]] 
  + plot_data[, names(plot_data)[j]]

This may not immediately result in your desired outcome, but you can use the same method to get the names for var.labels:

var.labels= c(names(plot_data)[i], names(plot_data)[j])

or whatever the exact syntax is. I don't have this package installed, but this should get you pretty close.