extrmadj: adjuste extreme values

In Nisus-Liu/bbturns: select turns form a time series by B-B algorithm

Description

Adjusting extreme values of series.

Usage

extrmadj(x, limit = 3.5, subs_method = c("spencer", "nearest"))

Arguments

x	numeric vector.
limit	threshold for recognizing extreme value.
subs_method	indicate method for substituting extreme values, spencer or nearest, the former means substituted by spencer value, latter means substituted by mean of nearest none extreme values.

Examples

##
usad1 = USAccDeaths
usad1[2:5] = mean(USAccDeaths) * 3  # first obs will become a extreme since the trend has been changed by 2:5
(ext1 = extrmadj(usad1))
plot(ext1$extr_infr$x, type = "l") # original
lines(ext1$x.nextr, type = "l", col = "red") # substituted
## 
usad2 = USAccDeaths
usad2[15] = - mean(USAccDeaths) * 3
(ext2 = extrmadj(usad2))
plot(ext2$extr_infr$x, type = "l") # original
lines(ext2$x.nextr, type = "l", col = "red") # substituted

## substituted by mean of nearest non-extrmadj
(ext3 = extrmadj(usad2, subs_method = "near"))
plot(ext3$extr_infr$x, type = "l") # original
lines(ext3$x.nextr, type = "l", col = "red") # substituted

Nisus-Liu/bbturns documentation built on May 7, 2019, 6:18 p.m.