winsor: Winsor normalization with fitting to <0,1> range.

In sidiropoulos/RNAprobR-old: An R package for analysis of massive parallel sequencing based RNA structure probing data

Description

Function performs Winsor normalization of a supplied vector. Steps: 1. Calcualate top winsor value [(1+winsor_level)/2 quantile], and bottom winsor value ((1-winsor_level)/2 quantile) 2. Each value below bottom winsor value set to bottom winsor value; each value above top winsor value set to top winsor value 3. Transform linearly all the values to [0,1] range

Usage

winsor(input_vector, winsor_level = 0.9, only_top = FALSE)

Arguments

input_vector	Vector with values to be Winsorized
winsor_level	Winsorization level. Bottom outliers will be set to (1-winsor_level)/2 quantile and top outliers to (1+winsor_level)/2 quantile.
only_top	If TRUE then bottom values are not Winsorized and the lowest is set to 0.

Value

Vector of numerics within <0,1>.

Author(s)

Lukasz Jan Kielpinski

References

Hastings, Cecil; Mosteller, Frederick; Tukey, John W.; Winsor, Charles P. Low Moments for Small Samples: A Comparative Study of Order Statistics. The Annals of Mathematical Statistics 18 (1947), no. 3, 413–426.

Examples

data_set <- runif(1:100)*100
plot(winsor(data_set, winsor_level=0.8) ~ data_set)

sidiropoulos/RNAprobR-old documentation built on May 29, 2019, 9:58 p.m.