imputation function

Missing data imputation (e.g. substitution by value or hotdeck method).

imputation(imethod = "value", D, Attribute = NULL, Missing = NA, Value = 1)

Arguments

• imethod: imputation method type:

  • value -- substitutes missing data by Value (with single element or several elements);
  • hotdeck -- searches first the most similar example (i.e. using a k-nearest neighbor method -- knn) in the dataset and replaces the missing data by the value found in such example;

• D: dataset with missing data (data.frame)

• Attribute: if NULL then all attributes (data columns) with missing data are replaced. Else, Attribute is the attribute number (numeric) or name (character).

• Missing: missing data symbol

• Value: the substitution value (if imethod=value) or number of neighbors (k of knn).

Details

Check the references.

Returns

A data.frame without missing data.

References

• M. Brown and J. Kros.

  Data mining and the impact of missing data.

  In Industrial Management & Data Systems, 103(8):611-621, 2003.

• This tutorial shows additional code examples:

  P. Cortez.

  A tutorial on using the rminer R package for data mining tasks.

  Teaching Report, Department of Information Systems, ALGORITMI Research Centre, Engineering School, University of Minho, Guimaraes, Portugal, July 2015.

  http://hdl.handle.net/1822/36210

Author(s)

Paulo Cortez http://www3.dsi.uminho.pt/pcortez/

See Also

fit and delevels.

Note

See also http://hdl.handle.net/1822/36210 and http://www3.dsi.uminho.pt/pcortez/rminer.html

Examples

d=matrix(ncol=5,nrow=5)
d[1,]=c(5,4,3,2,1)
d[2,]=c(4,3,4,3,4)
d[3,]=c(1,1,1,1,1)
d[4,]=c(4,NA,3,4,4)
d[5,]=c(5,NA,NA,2,1)
d=data.frame(d); d[,3]=factor(d[,3])
print(d)
print(imputation("value",d,3,Value="3"))
print(imputation("value",d,2,Value=median(na.omit(d[,2]))))
print(imputation("value",d,2,Value=c(1,2)))
print(imputation("hotdeck",d,"X2",Value=1))
print(imputation("hotdeck",d,Value=1))

## Not run:

# hotdeck 1-nearest neighbor substitution on a real dataset:
require(kknn)
d=read.table(
   file="http://archive.ics.uci.edu/ml/machine-learning-databases/autos/imports-85.data",
   sep=",",na.strings="?",stringsAsFactors=TRUE)
print(summary(d))
d2=imputation("hotdeck",d,Value=1)
print(summary(d2))
par(mfrow=c(2,1))
hist(d$V26)
hist(d2$V26)
par(mfrow=c(1,1)) # reset mfrow
## End(Not run)