[R] Change points in R

Achim Zeileis Achim.Zeileis at uibk.ac.at
Fri Jan 31 09:02:52 CET 2014


On Fri, 31 Jan 2014, Benjamin Ward (ENV) wrote:

> Hi R helpers,
>
> I have a set of data best shown in this below graph.
>
> Each coloured line represents a statistic calculated across pairs of DNA 
> sequences. And for each coloured line, I would like to identify 
> breakpoints - so identify the chunks where the values are high, for 
> example, in the light blue line, there is a large high segment just 
> after x=2e+05. From googling the aim to find such points, I've read 
> about something called change-point analysis, used with time series data 
> and I wondered if it or a variant of it in R might be of use here, this 
> data is a series of % values (double), all a single measurement i.e. for 
> each line, a 'scanner' passed over two sequences and at each step 
> recorded the % value. Can change-point analysis help me here and if so 
> what package or method will allow me to do this making as little 
> assumptions about my data as possible?

The graph didn't make it through but from what you describe it seems that 
the "tilingArray" package on Bioconductor would be helpful for you. See 
als Huber et al. (2006, Bioinformatics, 22(16), 1963-1970).

Other useful packages on CRAN include the packages: bcp, changepoint, cpm, 
segmented and strucchange (among others).

> Thanks in advance,
>
> Ben W.
>
> [X]
>
> 	[[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>




More information about the R-help mailing list