dlookr (version 0.5.0)

binning: Binning the Numeric Data

Description

The binning() converts a numeric variable to a categorization variable.

Usage

binning(
  x,
  nbins,
  type = c("quantile", "equal", "pretty", "kmeans", "bclust"),
  ordered = TRUE,
  labels = NULL,
  approxy.lab = TRUE
)

Arguments

x

numeric. numeric vector for binning.

nbins

integer. number of intervals(bins). required. if missing, nclass.Sturges is used.

type

character. binning method. Choose from "quantile", "equal", "equal", "pretty", "kmeans" and "bclust". The "quantile" sets breaks with quantiles of the same interval. The "equal" sets breaks at the same interval. The "pretty" chooses a number of breaks not necessarily equal to nbins using base::pretty function. The "kmeans" uses stats::kmeans function to generate the breaks. The "bclust" uses e1071::bclust function to generate the breaks using bagged clustering. "kmeans" and "bclust" was implemented by classInt::classIntervals() function.

ordered

logical. whether to build an ordered factor or not.

labels

character. the label names to use for each of the bins.

approxy.lab

logical. If TRUE, large number breaks are approximated to pretty numbers. If FALSE, the original breaks obtained by type are used.

Value

An object of bins class. Attributes of bins class is as follows.

• type : binning type, "quantile", "equal", "pretty", "kmeans", "bclust".

• breaks : the number of intervals into which x is to be cut.

• levels : levels of binned value.

• raw : raw data, numeric vector corresponding to x argument.

"bins" class attributes information

Attributes of the "bins" class that is as follows.

• class : "bins"

• levels : levels of factor or ordered factor

• type : binning method

• breaks : breaks for binning

• raw : raw data before binning

See vignette("transformation") for an introduction to these concepts.

Details

This function is useful when used with the mutate/transmute function of the dplyr package.

See Also

binning_by, print.bins, summary.bins, plot.bins.

Examples

# NOT RUN {
# Generate data for the example
heartfailure2 <- heartfailure
heartfailure2[sample(seq(NROW(heartfailure2)), 20), "platelets"] <- NA

# Binning the carat variable. default type argument is "quantile"
bin <- binning(heartfailure2$platelets)
# Print bins class object
bin

# Summarise bins class object
summary(bin)

# Plot bins class object
plot(bin)

# Using labels argument
bin <- binning(heartfailure2$platelets, nbins = 4,
              labels = c("LQ1", "UQ1", "LQ3", "UQ3"))
bin
# Using another type argument
bin <- binning(heartfailure2$platelets, nbins = 5, type = "equal")
bin
# bin <- binning(heartfailure2$platelets, nbins = 5, type = "pretty")
# bin
# bin <- binning(heartfailure2$platelets, nbins = 5, type = "kmeans")
# bin
# bin <- binning(heartfailure2$platelets, nbins = 5, type = "bclust")
# bin

x <- sample(1:1000, size = 50) * 12345679
bin <- binning(x)
bin
bin <- binning(x, approxy.lab = FALSE)
bin

# extract binned results
extract(bin)

# -------------------------
# Using pipes & dplyr
# -------------------------
library(dplyr)

# Compare binned frequency by death_event
heartfailure2 %>%
 mutate(platelets_bin = binning(heartfailure2$platelets) %>% 
                     extract()) %>%
 group_by(death_event, platelets_bin) %>%
 summarise(freq = n()) %>%
 arrange(desc(freq)) %>%
 head(10)
 
 # Compare binned frequency by death_event using Viz
 heartfailure2 %>%
   mutate(platelets_bin = binning(heartfailure2$platelets) %>% 
           extract()) %>%
   target_by(death_event) %>% 
   relate(platelets_bin) %>% 
   plot()
 
# }