Replace x-axis label from the plot of a MarkovChain object

Question

Here is some code to generate a 0th order Markov Chain plot. I'd like to replace the plot's x-axis labels (c, d, h, i, o, p) with 45 degree rotated labels showing the first six months of the year. However, using xaxt="n" in the plot call doesn't seem to work. This code just writes over the existing labels, rather than replacing them. How can I replace the labels with the ones I want?

library(clickstream)
clickstreams <- c("User1,h,c,c,p,c,h,c,p,p,c,p,p,o",
                   "User2,i,c,i,c,c,c,d",
                   "User3,h,i,c,i,c,p,c,c,p,c,c,i,d",
                   "User4,c,c,p,c,d",
                   "User5,h,c,c,p,p,c,p,p,p,i,p,o",
                   "User6,i,h,c,c,p,p,c,p,c,d")
csf <- tempfile()
writeLines(clickstreams, csf)
cls <- readClickstreams(csf, header = TRUE)
mc <- fitMarkovChain(cls, order=0)
plot(mc, xaxt="n")
text(x=1:6, y=par()$usr[3], labels = month.name[1:6], srt=45, adj = c(1.1,1.1), xpd = TRUE, cex=.9)

score 0 · Answer 1 · answered Apr 09 '19 at 15:19

0

Using plot.default rather than plot exposes the graphics plot function, which allows the use of the xaxt function parameter to remove the existing x-axis labels. Since plot.default won't accept a MarkovChain object, the x and y values for the plot need to be extracted from the object.

plot.default(
   x=mc@transitions[[1]]$states, 
   y=mc@transitions[[1]]$probability,
   xaxt="n", 
   ann=FALSE, 
   pch="-", 
   cex=3
)

text(
    x = 1:6, 
    y = par()$usr[3], 
    labels = month.name[1:6], 
    srt = 45, 
    adj = c(1.1,1.1), 
    xpd = TRUE, 
    cex = .9
)

answered Apr 09 '19 at 15:19

meenaparam

1,949
2
17
29

1

`with(mc@transitions[[1]], plot(states, probability, xaxt="n"))` would also work. – jay.sf Apr 09 '19 at 15:30
1

Thanks, your answer is better than mine as it uses the original styling for the clickstream plot function. If you want to submit it as an answer, I'll accept it. – meenaparam Apr 09 '19 at 16:11
Done so, glad to help! – jay.sf Apr 09 '19 at 16:49

score 0 · Accepted Answer · answered Apr 09 '19 at 16:49

str(mc) reveals the structure of the S4 object. Passing into plot just the xy-coordinates rather than the whole "MarkovChain" object recovers the xaxt option functionality. Using with makes this most convenient.

Moreover this allows us to use the implemented S4 plot method for signature 'MarkovChain' of the clickstream package.

library(clickstream)
with(mc@transitions[[1]], plot(states, probability, xaxt="n"))
text(x=1:6, y=par()$usr[3], labels=month.name[1:6], srt=45, adj=rep(1.1, 2), xpd=TRUE, cex=.9)

Result

Data

library(clickstream)
mc <- new("MarkovChain", states = c("h", "c", "p", "o", "i", "d"), 
    order = 0, transitions = list(structure(list(states = structure(1:6, .Label = c("c", 
    "d", "h", "i", "o", "p"), class = "factor"), frequency = c(25L, 
    4L, 5L, 7L, 2L, 17L), probability = c(0.416666666666667, 
    0.0666666666666667, 0.0833333333333333, 0.116666666666667, 
    0.0333333333333333, 0.283333333333333)), class = "data.frame", row.names = c(NA, 
    -6L))), lambda = 0, logLikelihood = -88.4241188515082, observations = 60, 
    start = structure(c(c = 0.166666666666667, h = 0.5, i = 0.333333333333333
    ), class = "table", .Dim = 3L, .Dimnames = structure(list(
        c("c", "h", "i")), .Names = "")), end = structure(c(d = 0.666666666666667, 
    o = 0.333333333333333), class = "table", .Dim = 2L, .Dimnames = structure(list(
        c("d", "o")), .Names = "")), transientStates = c("c", 
    "h", "i", "p"), absorbingStates = c("d", "o"), absorbingProbabilities = structure(list(
        state = structure(1:4, .Label = c("c", "h", "i", "p"), class = "factor"), 
        d = c(0.792201957232916, 0.864224245314581, 0.903842865008182, 
        0.517925028202586), o = c(0.207798042767084, 0.13577575468542, 
        0.0961571349918185, 0.482074971797414)), class = "data.frame", row.names = c(c = 1L, 
    h = 3L, i = 4L, p = 6L)))

Replace x-axis label from the plot of a MarkovChain object

2 Answers2