Winsorizes the proportion of ordered data given by lambda from each tail.
Usage
win(x, lambda = 0.2)
Value
Returns Winsorized data.
Arguments
x
A vector of data.
lambda
A proportion from 0-1 giving the amount of data to be Winsorized in each tail of an ordered dataset.
Author
Ken Aho
Details
In Winsorization we we replace responses that are not in the central \(1 - 2\lambda\) part of an ordered sample with the minimum and maximum responses of the central part of the sample.
References
Wilcox, R. R. (2005) Introduction to Robust Estimation and Hypothesis Testing, Second
Edition. Elsevier, Burlington, MA.