Equal Frequency Discretization in R

Equal frequency discretization in R

EDIT : given your real goal, why don't you just do (corrected) :

 EqualFreq2 <- function(x,n){
    nx <- length(x)
    nrepl <- floor(nx/n)
    nplus <- sample(1:n,nx - nrepl*n)
    nrep <- rep(nrepl,n)
    nrep[nplus] <- nrepl+1
    x[order(x)] <- rep(seq.int(n),nrep)
    x
}

This returns a vector with indicators for which bin they are. But as some values might be present in both bins, you can't possibly define the bin limits. But you can do :

x <- rpois(50,5)
y <- EqualFreq2(x,15)
table(y)
split(x,y)

---

Original answer:

You can easily just use cut() for this :

EqualFreq <-function(x,n,include.lowest=TRUE,...){
    nx <- length(x)    
    id <- round(c(1,(1:(n-1))*(nx/n),nx))

    breaks <- sort(x)[id]
    if( sum(duplicated(breaks))>0 stop("n is too large.")

    cut(x,breaks,include.lowest=include.lowest,...)

}

Which gives :

set.seed(12345)
x <- rnorm(50)
table(EqualFreq(x,5))

 [-2.38,-0.886] (-0.886,-0.116]  (-0.116,0.586]   (0.586,0.937]     (0.937,2.2] 
             10              10              10              10              10 

x <- rpois(50,5)
table(EqualFreq(x,5))

 [1,3]  (3,5]  (5,6]  (6,7] (7,11] 
    10     13     11      6     10 

As you see, for discrete data an optimal equal binning is rather impossible in most cases, but this method gives you the best possible binning available.

Equal frequency and equal width binning in R

For equal width binning, I suggest using the classInt package:

dataset <- c(0, 4, 12, 16, 16, 18, 24, 26, 28)

library(classInt)
classIntervals(dataset, 4)
x <- classIntervals(dataset, 4, style = 'equal')

To use the breaks, you can check x$brks.

As for equal frequency binning, you could use the same package, with option style = 'quantile':

classIntervals(dataset, 4, style = 'quantile')

It doesn't separate in exactly equal sized bins because of repeated values in dataset (16) and because the dataset simply can't be split exactly in 4 bins with strictly the same number of elements because it has 9 elements. I don't know if this is an issue, because in the reference provided, it says that

"... each group contains approximately same number of values."

As you didn't explicit the exact method you are looking for, I suggest referring to this post for another method, in your example it would be:

library(Hmisc)
table(cut2(dataset, m = length(dataset)/4))

Also, the other posts in the link suggested above provide other alternatives and some relevant discussion about these methods.

removing noises with equal depth binning in R by replacing each bins with its mean or median

We can create groups by dividing length of a in equal number of bins and use ave to calculate rounded mean in each group.

no_of_bins <- 4
round(ave(a, rep(1:length(a), each = no_of_bins, length.out = length(a))))
#[1]  9  9  9  9 23 23 23 23 29 29 29 29

PS -

• ave has default function as mean so it has not been explicitly applied.

Discretization in R

You can still use the dprep package, but you have to install it from source (I just tested and it works well). However, you may well have a look at the discretization or infotheo packages which provide similar functionalities, e.g. equal interval width, equal frequency intervals, ChiMerge, etc.

Splitting a continuous variable into equal sized groups

try this:

split(das, cut(das$anim, 3))

if you want to split based on the value of wt, then

library(Hmisc) # cut2
split(das, cut2(das$wt, g=3))

anyway, you can do that by combining cut, cut2 and split.

UPDATED

if you want a group index as an additional column, then

das$group <- cut(das$anim, 3)

if the column should be index like 1, 2, ..., then

das$group <- as.numeric(cut(das$anim, 3))

UPDATED AGAIN

try this:

> das$wt2 <- as.numeric(cut2(das$wt, g=3))
> das
   anim    wt wt2
1     1 181.0   1
2     2 179.0   1
3     3 180.5   1
4     4 201.0   2
5     5 201.5   2
6     6 245.0   2
7     7 246.4   3
8     8 189.3   1
9     9 301.0   3
10   10 354.0   3
11   11 369.0   3
12   12 205.0   2
13   13 199.0   1
14   14 394.0   3
15   15 231.3   2

R plot frequency distribution

You could plot the table of your data:

xx<-sample(1:10,100,replace=TRUE)
plot(table(xx))

---

Related Topics