midp2brk {lazy.tools} | R Documentation |
Define Break Points from Midpoints
midp2brk(midpoints, min = NULL, max = NULL, margin = NULL)
midpoints |
a vector of midpoints. |
min |
value of the lowest break. |
max |
value of the highest break. |
margin |
value to be subtracted/added to
the lowest and highest midpoints. |
Break values from midpoint values:
brk[1]=midp[1] - margin
brk[2]=midp[1] + (midp[2]-midp[1])/2
brk[3]=midp[2] + (midp[3]-midp[2])/2
brk[length(midpoints)+1]=midp[length(midpoints)] + margin
where margin=(midp[2]-midp[1])/2
.
In general, the break point is located at the center of two midpoints:
brk[k]=0.5*(midp[k-1]+midp[k])
Midpoint values from break values:
midp[1]=0.5*(brk[1]+brk[2])
or midp[1]=min
midp[2]=0.5*(brk[2]+brk[3])
midp[3]=2*brk[3]-midp[2]
brk[length(brk)-1]=2*brk[length(brk)-1]-midp[length(brk)-2]
Note that midp[1] is treated separately and it is assumed that
the break point is located at the center of two midpoints.
Try below to see how the obs on the border are classified.
xx <- c(1,2,3,4,5,6,7)
brks <- c(0,2,4,6,8)
cc1 <- cut(xx,brks)
cc2 <- cut(xx,brks, right=FALSE)
Print(xx,as.character(cc1),as.character(cc2))
A vector of break points of length length(midpoints)+1.
midp <- seq(0,5,length.out=11) brk <- midp2brk( midp, margin=1 ) Print(midp,brk) midp <- seq(0,5,length.out=11) brk <- midp2brk( midp, margin=1 ) Print(midp,brk) set.seed(1701) x <- rnorm(100)+2.5 # histogram by native hist resh <- hist( x, breaks=brk ) # histogram by native barplot # freq <- resh$counts # names(freq) <- midp # barplot( freq, space=0 ) barplot( resh$count,names.arg=midp, space=0, xlab="X", col="white" , sub="top and bottom bars contain outliers." ) # Below causes error because 'brk' does not span the range of 'x'. # brk <- midp2brk( midp ) # resh <- hist( x, breaks=brk )