[R-sig-eco] help

[Sarah Goslee]

Hi Mahnaz,

Please send your replies to the mailing list, not just to me.

I'm still not sure I understand where your problem lies.

You do not need to transform the data prior to NMDS.

If you want help with the "plot being empty" we need to know basic
details about your data, the commands you used, etc. - ideally a
reproducible example.

Sarah

On Sat, Dec 10, 2011 at 12:01 AM, Mahnaz Rabbaniha <rab.mahnaz at gmail.com> wrote:
>  Dear Sarah
>
>
> thank you for your attention,
>
> I want to use PCA with supplementary variables due to correlation
> between the independent factor such as salinity ,temperature... with
> dependent variable ( fish larva ) therefore changed  dependent
> variable with log(x+1) and tested normality with shapiro test but data
> haven't changed and remained non normality then i decided to used the
> nonparametric analysis MDS   and found its order from net but it
> wasn't successfully and plot of matrix was empty.
>
> my question is: for using the MDS have to need the special form of
> excel , i have to use from ternsformed data,
> I will be happy to provide any kind of solution
>
> thanks
>
> mahnaz
>
>
> On 12/9/11, Sarah Goslee <sarah.goslee at gmail.com> wrote:
>> Hi,
>>
>> On Fri, Dec 9, 2011 at 7:13 AM, Mahnaz Rabbaniha <rab.mahnaz at gmail.com>
>> wrote:
>>> Hi
>>>
>>> in my project i collected its data in excel sheet that i attached,data
>>> were non normal then i transformed them and check the normality by
>>> shapiro.test that it showed again non-normality therefore i decided
>>> used the MDS analyses in r and download this site
>>>  ( http://www.statmethods.net/advstats/mds.html)
>>>
>>> but i haven't done it,  i think  the main problem is the form of excel
>>> matrix for doing it and also i want to know about different between
>>>  CLASSICAL
>>> MDS and isoMDS.
>>
>> We need to know more about what you're donig to be able to help with
>> the excel matrix. What is the problem? How are you importing it into R?
>>
>> ?cmdscale
>> library(MASS)
>> ?isoMDS
>>
>> offers some information about the algorithms used in the two approaches,
>> as well as many references for more information
>>
>> Briefly and simplistically, the difference has to do with whether the
>> algorithm
>> attempts to preserve actual dissimilarities or rank-order dissimilarities.
>>
>>> please help me because this method is new in iran
>>>
>>> Thanks in advance
>>> --
>>> *Mahnaz Rabbaniha*
>

-- 
Sarah Goslee
http://www.functionaldiversity.org