R help - Dec 2011 - How to automate the detection of break points for use in cut

Dear R-users,

I would like to know if there is a function (in base R or the extension
packages) that would automatically detect the break points in a vector x
for later use in the cut function. The idea is to determine the boundaries
of the n intervals (n>=1) delimiting clusters of data points which could be
considered "reasonably" close, given a numerical vector x with unknown
content and unknown multimodal distribution.

For instance, given for the vector x defined by set.seed(1234); x <-
sort(c(rnorm(20,-1,0.1),rnorm(
10,5,0.1),rnorm(10,100,0.1))), this function would return a vector of 4
points: min(x), one value between 20 and 5, one value between 5 and 100,
and max(x).

Thank you in advance for your suggestions.

Sebastien

	[[alternative HTML version deleted]]

Obviously, cut would do the job if one knows the number of intervals in
advance, which I assume I won't. I guess what I'm looking for is a
function
that figures out the number of intervals and their boundaries.

Sebastien

On Tue, Dec 6, 2011 at 3:29 AM, Sébastien Bihorel <pomchip@free.fr> wrote:
> Dear R-users,
>
> I would like to know if there is a function (in base R or the extension
> packages) that would automatically detect the break points in a vector x
> for later use in the cut function. The idea is to determine the boundaries
> of the n intervals (n>=1) delimiting clusters of data points which could
be
> considered "reasonably" close, given a numerical vector x with
unknown
> content and unknown multimodal distribution.
>
> For instance, given for the vector x defined by set.seed(1234); x <-
> sort(c(rnorm(20,-1,0.1),rnorm(
> 10,5,0.1),rnorm(10,100,0.1))), this function would return a vector of 4
> points: min(x), one value between 20 and 5, one value between 5 and 100,
> and max(x).
>
> Thank you in advance for your suggestions.
>
> Sebastien
>
	[[alternative HTML version deleted]]

I forgot to post my reply to Duncan's response on the list.

On Tue, Dec 6, 2011 at 7:56 AM, Sébastien Bihorel <pomchip@free.fr> wrote:
> Thanks for the link Duncan,
>
> Given the number of methods and links listed in the Cluster task view,
> things are looking a bit more complex than I thought... I'll have to
read
> more about clustering before I can start testing.
>
> Sebastien
>
>
> On Tue, Dec 6, 2011 at 7:28 AM, Duncan Murdoch
<murdoch.duncan@gmail.com>wrote:
>
>> On 11-12-06 3:34 AM, Sébastien Bihorel wrote:
>>
>>> Obviously, cut would do the job if one knows the number of
intervals in
>>> advance, which I assume I won't. I guess what I'm looking
for is a
>>> function
>>> that figures out the number of intervals and their boundaries.
>>>
>>
>> That's not really a simple problem, but there are functions that do
>> clustering and fit mixture models to data, which might be close enough.
>>  See the Cluster task view at http://cran.r-project.org/web/**
>> views/Cluster.html
<http://cran.r-project.org/web/views/Cluster.html>.
>>
>> Duncan Murdoch
>>
>>
>>> Sebastien
>>>
>>> On Tue, Dec 6, 2011 at 3:29 AM, Sébastien
Bihorel<pomchip@free.fr>
>>>  wrote:
>>>
>>>  Dear R-users,
>>>>
>>>> I would like to know if there is a function (in base R or the
extension
>>>> packages) that would automatically detect the break points in a
vector x
>>>> for later use in the cut function. The idea is to determine the
>>>> boundaries
>>>> of the n intervals (n>=1) delimiting clusters of data points
which
>>>> could be
>>>> considered "reasonably" close, given a numerical
vector x with unknown
>>>> content and unknown multimodal distribution.
>>>>
>>>> For instance, given for the vector x defined by set.seed(1234);
x<-
>>>> sort(c(rnorm(20,-1,0.1),rnorm(
>>>> 10,5,0.1),rnorm(10,100,0.1))), this function would return a
vector of 4
>>>> points: min(x), one value between 20 and 5, one value between 5
and 100,
>>>> and max(x).
>>>>
>>>> Thank you in advance for your suggestions.
>>>>
>>>> Sebastien
>>>>
>>>>
>>>        [[alternative HTML version deleted]]
>>>
>>>
>>>
>>>
>>> ______________________________**________________
>>> R-help@r-project.org mailing list
>>>
https://stat.ethz.ch/mailman/**listinfo/r-help<https://stat.ethz.ch/mailman/listinfo/r-help>
>>> PLEASE do read the posting guide http://www.R-project.org/**
>>> posting-guide.html
<http://www.R-project.org/posting-guide.html>
>>> and provide commented, minimal, self-contained, reproducible code.
>>>
>>
>>
>
	[[alternative HTML version deleted]]