Scatter Plot Matrices

Question

I have a matrix mat[n,m] and I'd like to use splom to plot the scatterplots of mat[,"col4"] as a function of all other column values. Also I'd like to add different colors to points of certain row numbers which are stored in rownID[] . I've seen examples using splom but they plot all variables against all variables and use group of columns to change the color of points. Is it possible to do what I want using splom (or other R function)?

Example:

set.seed(1)
mat <-  matrix(sample(0:100, 16), ncol=4)

dimnames(mat) <- list(rownames(mat, do.NULL = FALSE, prefix = "row"),
                          colnames(mat, do.NULL = FALSE, prefix = "col"))

mat
     col1 col2 col3 col4
row1   26   19   58   61
row2   37   86    5   33
row3   56   97   18   66
row4   89   62   15   42

rowID <- matrix(c(1,3), ncol=1, nrow=2)

Thanks to https://stackoverflow.com/a/16033003/1262767

I've been using featurePlot function of caret package but I don't know how to change color of some specific points (that's why I'm interested in splom ):

featurePlot(mat, mat$col4, plot = "scatter",
            ## Add some space between the panels
            between = list(x = 1, y = 1), main = "testSet",
            ## Add a background grid ('g') and a smoother ('smooth')
            type = c("g", "p", "s"))

Answer 1

This doesn't really seem like a good fit for splom . I think you're be better off reshaping your data and using a standard xyplot . For example

library(reshape2)
mm<-melt(cbind(data.frame(mat), high=1:nrow(mat) %in% rowID), c("col4","high"))
xyplot(col4~value|variable, mm, groups=high)

which gives