R Graphical Manual

Browse All

Last data update: 2014.03.03

R: Residency analysis using a changepoint model

changeLight

R Documentation

Residency analysis using a changepoint model

Description

Function to discriminate between periods of residency and movement based on consecutive sunrise and sunset data. The calculation is based on a changepoint model (R Package changepoint: cpt.mean) to find multiple changepoints within the data.

Usage

changeLight(tFirst, tSecond, type, twl, quantile = 0.9, rise.prob = NA,
  set.prob = NA, days = 5, plot = TRUE, summary = TRUE)

Arguments

`tFirst`	vector of sunrise/sunset times (e.g. 2008-12-01 08:30).
`tSecond`	vector of of sunrise/sunset times (e.g. 2008-12-01 17:30).
`type`	vector of either 1 or 2, defining `tFirst` as sunrise or sunset respectively.
`twl`	data.frame containing twilights and at least `tFirst`, `tSecond` and `type` (alternatively give each parameter separately).
`quantile`	probability threshold for stationary site selection. Higher values (above the defined quantile of all probabilities) will be considered as changes in the behavior. Argmuent will only be considered if either `rise.prob` and/or `set.prob` remain unspecified.
`rise.prob`	the probability threshold for sunrise: greater or equal values indicates changes in the behaviour of the individual.
`set.prob`	the probability threshold for sunset: higher and equal values indicates changes in the behaviour of the individual.
`days`	a threshold for the length of stationary period. Periods smaller than "days" will not be considered as a residency period
`plot`	logical, if `TRUE` a plot will be produced
`summary`	logical, if `TRUE` a summary of the results will be printed

Details

The cpt.mean from the R Package changepoint is a function to find a multiple changes in mean for data where no assumption is made on their distribution. The value returned is the result of finding the optimal location of up to Q changepoints (in this case as many as possible) using the cumulative sums test statistic.

Value

A list with probabilities for sunrise and sunset the user settings of the probabilities and the resulting stationary periods given as a vector, with the residency sites as positiv numbers in ascending order (0 indicate movement/migration).

Note

The sunrise and/or sunset times shown in the graph (if plot=TRUE) represent hours of the day. However if one or both of the twilight events cross midnight during the recording period the values will be formed to avoid discontinuity.

Author(s)

Simeon Lisovski & Tamara Emmenegger

References

Taylor, Wayne A. (2000) Change-Point Analysis: A Powerful New Tool For Detecting Changes.

M. Csorgo, L. Horvath (1997) Limit Theorems in Change-Point Analysis. Wiley.

Chen, J. and Gupta, A. K. (2000) Parametric statistical change point analysis. Birkhauser.

Examples

data(hoopoe2)
  hoopoe2$tFirst <- as.POSIXct(hoopoe2$tFirst, tz = "GMT")
  hoopoe2$tSecond <- as.POSIXct(hoopoe2$tSecond, tz = "GMT")
residency <- changeLight(hoopoe2, quantile=0.9)

Results


R version 3.3.1 (2016-06-21) -- "Bug in Your Hair"
Copyright (C) 2016 The R Foundation for Statistical Computing
Platform: x86_64-pc-linux-gnu (64-bit)

R is free software and comes with ABSOLUTELY NO WARRANTY.
You are welcome to redistribute it under certain conditions.
Type 'license()' or 'licence()' for distribution details.

R is a collaborative project with many contributors.
Type 'contributors()' for more information and
'citation()' on how to cite R or R packages in publications.

Type 'demo()' for some demos, 'help()' for on-line help, or
'help.start()' for an HTML browser interface to help.
Type 'q()' to quit R.

> library(GeoLight)
Loading required package: maps

 # maps v3.1: updated 'world': all lakes moved to separate new #
 # 'lakes' database. Type '?world' or 'news(package="maps")'.  #


> png(filename="/home/ddbj/snapshot/RGM3/R_CC/result/GeoLight/changeLight.Rd_%03d_medium.png", width=480, height=480)
> ### Name: changeLight
> ### Title: Residency analysis using a changepoint model
> ### Aliases: changeLight
> 
> ### ** Examples
> 
> data(hoopoe2)
>   hoopoe2$tFirst <- as.POSIXct(hoopoe2$tFirst, tz = "GMT")
>   hoopoe2$tSecond <- as.POSIXct(hoopoe2$tSecond, tz = "GMT")
> residency <- changeLight(hoopoe2, quantile=0.9)

Probability threshold(s):
 
	Sunrise:  0.03757	Sunset:  0.01307
 
 
Migration schedule table:
 
  Site             Arrival           Departure Days     P.start       P.end
1    a 2008-07-15 23:34:00 2008-07-24 23:32:30  9.0 0.010416667 0.008333333
2    b 2008-07-31 00:15:00 2008-08-19 12:26:00 19.5 0.010500257 0.000000000
3    c 2008-08-25 00:25:30 2008-09-01 12:20:00  7.5 0.008333333 0.006995885
4    d 2008-09-12 12:27:30 2008-10-02 12:28:00 20.0 0.006321499 0.000000000
5    e 2008-10-03 12:22:30 2008-10-19 12:26:00 16.0 0.004629630 0.000000000
6    f 2008-10-20 12:22:00 2008-12-03 12:35:30 44.0 0.000000000 0.006198347
7    g 2008-12-04 12:28:30 2008-12-18 12:32:00 14.0 0.000000000 0.010706019
8    h 2008-12-19 12:41:00                <NA>   NA 0.005729167 0.000000000
  Days.1   P.start.1
1    9.0 0.010416667
2   19.5 0.010500257
3    7.5 0.008333333
4   20.0 0.006321499
5   16.0 0.004629630
6   44.0 0.000000000
7   14.0 0.000000000
8     NA 0.005729167
> 
> 
> 
> 
> 
> dev.off()
null device 
          1 
>