I have data that looks like this:
#val Freq1 Freq2
0.000 178 202
0.001 4611 5300
0.002 99 112
0.003 26 30
0.004 17 20
0.005 15 20
0.006 11 14
0.007 11 13
0.008 13 13
...many more lines..
Full data can be found here: http://dpaste.com/173536/plain/
What I intend to do is to have a cumul开发者_运维问答ative graph with "val" as x-axis with "Freq1" & "Freq2" as y-axis, plot together in 1 graph.
I have this code. But it creates two plots instead of 1.
dat <- read.table("stat.txt",header=F);
val<-dat$V1
freq1<-dat$V2
freq2<-dat$V3
valf1<-rep(val,freq1)
valf2<-rep(val,freq2)
valfreq1table<- table(valf1)
valfreq2table<- table(valf2)
cumfreq1=c(0,cumsum(valfreq1table))
cumfreq2=c(0,cumsum(valfreq2table))
plot(cumfreq1, ylab="CumFreq",xlab="Loglik Ratio")
lines(cumfreq1)
plot(cumfreq2, ylab="CumFreq",xlab="Loglik Ratio")
lines(cumfreq2)
What's the right way to approach this?
data <- read.table("http://dpaste.com/173536/plain/", header = FALSE)
sample1 <- unlist(apply(as.matrix(data),1,function(x) rep(x[1],x[2])))
sample2 <- unlist(apply(as.matrix(data),1,function(x) rep(x[1],x[3])))
plot(ecdf(sample1), verticals=TRUE, do.p=FALSE,
main="ECDF plot for both samples", xlab="Scores",
ylab="Cumulative Percent",lty="dashed")
lines(ecdf(sample2), verticals=TRUE, do.p=FALSE,
col.h="red", col.v="red",lty="dotted")
legend(100,.8,c("Sample 1","Sample 2"),
col=c("black","red"),lty=c("dashed","dotted"))
Try the ecdf()
function in base R --- which uses plot.stepfun()
if memory serves --- or the Ecdf()
function in Hmisc by Frank Harrell. Here is an example from help(Ecdf)
that uses a grouping variable to show two ecdfs in one plot:
# Example showing how to draw multiple ECDFs from paired data
pre.test <- rnorm(100,50,10)
post.test <- rnorm(100,55,10)
x <- c(pre.test, post.test)
g <- c(rep('Pre',length(pre.test)),rep('Post',length(post.test)))
Ecdf(x, group=g, xlab='Test Results', label.curves=list(keys=1:2))
Just for the record, here is how you get multiple lines in the same plot "by hand":
plot(cumfreq1, ylab="CumFreq",xlab="Loglik Ratio", type="l")
# or type="b" for lines and points
lines(cumfreq2, col="red")
精彩评论