quantile.fdt: Quantile of frequency distribution table (numerical variable)

In fdth: Frequency Distribution Tables, Histograms and Polygons

Quantile of frequency distribution table (numerical variable)

Description

S3 methods for the quantile of a fdt.
Useful to estimate the quantile (when the real data vector is not known) from a previous fdt.

Usage

## S3 methods: numerical
## S3 method for class 'fdt'
quantile(x,
         ...,
         i=1,
         probs=seq(0, 1, 0.25))

## S3 method for class 'fdt.multiple'
quantile(x, ...) 

Arguments

x	a fdt (simple or multiple) object.
i	an integer vector indicating which quantiles should be computed. Values must be in 1:(length(probs)-1).
probs	a finite numeric vector in [0, 1] defining the quantile scale. It must have at least two values and be sorted in non-decreasing order.
...	potential further arguments (required by generic).

Details

quantile.fdt calculates the quantiles based on a known formula for class intervals. quantile.fdt.multiple calls quantile.fdt for each variable, that is, each column of the data.frame.

Value

quantile.fdt returns a named numeric vector containing the value(s) of the quantile(s) from fdt. Names are derived from the selected probability levels in probs (for example, 25%, 50%, 75%). quantile.fdt.multiple returns a list, where each element is a numeric vector containing the quantile(s) of the fdt for each variable.

Author(s)

Faria, J. C.
Allaman, I. B
Jelihovschi, E. G.

See Also

median.fdt, var.fdt, mfv.

Examples

library(fdth)

x <- rnorm(n=1e3,
           mean=5,
           sd=1)

(ft <- fdt(x))

# Quartile from vector
quantile(x)[2:4]

# Quartile from grouped data in a fdt object
quantile(ft,
         i=1:3)

# Quartile from vector
quantile(x,
         probs=seq(from=0,
                   to=1,
                   by=.1))[2:10]

# Decile from grouped data in a fdt object
quantile(ft,
         i=1:9,
         probs=seq(from=0,
                   to=1,
                   by=.1))

# Percentile from vector
quantile(x,
         probs=seq(from=0,
                   to=1,
                   by=.01))[2:100]

# Percentile from grouped data in a fdt object
quantile(ft,
         i=1:99,
         probs=seq(from=0,
                   to=1,
                   by=.01))

# From a data.frame
mdf <- data.frame(x=rnorm(1e2, 
                          20,
                          2),
                  y=rnorm(1e2, 
                          30,
                          3),
                  z=rnorm(1e2,
                          40,
                          4))

head(mdf)

# From a data.frame (rows 2:4 are the 25%, 50%, and 75% quantiles)
apply(mdf,
      2,
      quantile)[2:4, ]

# From a fdt object
quantile(fdt(mdf),
         i=1:3)

## A small (but didactic) joke  
quantile(fdt(mdf),
         i=2,
         probs=seq(0, 
                   1, 
                   0.25))        # The quartile 2
quantile(fdt(mdf),
         i=5,
         probs=seq(0, 
                   1, 
                   0.10))        # The decile 5 

quantile(fdt(mdf),
         i=50,
         probs=seq(0, 
                   1, 
                   0.01))        # The percentile 50

quantile(fdt(mdf),
         i=500,
         probs=seq(0, 
                   1, 
                   0.001))       # The permile 500

median(fdt(mdf))                 # The median (all results are the same) ;)

# More than one quantile
quantile(fdt(mdf$x),
         i=1:3,
         probs=seq(0,
                   1,
                   0.25))  # The three quartiles

quantile(fdt(mdf$x),
         i=1:9,
         probs=seq(0,
                   1,
                   0.10))  # The nine deciles

# Legacy approach (no longer necessary)
# ql <- numeric()
# 
# for(i in 1:3)
#   ql[i] <- quantile(fdt(mdf$x),
#                     i=i,
#                     probs=seq(0,
#                               1,
#                               0.25))  # The three quartiles
# 
# names(ql) <- paste0(c(25,
#                       50,
#                       75),
#                     '%')
# round(ql,
#       2)

fdth documentation built on May 12, 2026, 1:08 a.m.