OTjitter: Add a small amount of noise to a numeric vector

In Renext: Renewal Method for Extreme Values Extrapolation

Add a small amount of noise to a numeric vector

Description

Add a small amount of noise to a numeric vector keeping all the values above the given threshold.

Usage

OTjitter(x, threshold = NULL)

Arguments

x	The numeric vector to which jitter should be added.
threshold	A threshold above which all elements of the modified vector must stay.

Value

A vector with the same length and nearly the same values as x. As in jitter, a small amount of noise is added to each value of x. The noise level is adjusted so that every noisy value remains above the specified threshold. When the a value is very close to the threshold, only a very small amount of negative noise can be added.

Note

The aim of this function is to remove possible ties in experimental OT data. Ties cause problems or warnings in some goodness-of-fit tests such as Kolmogorov-Smirnov.

Author(s)

Yves Deville

See Also

jitter

Examples

## Garonne data (heavily rounded)
x <- Garonne$OTdata$Flow
min(x) 
xmod <- OTjitter(x, threshold = 2500)
length(x)
nlevels(as.factor(x))
nlevels(as.factor(xmod))
max(abs(x-xmod))

Renext documentation built on Aug. 30, 2023, 1:06 a.m.