[R] nMDS with R: missing values

David L Carlson dc@r|@on @end|ng |rom t@mu@edu
Wed Apr 18 15:58:05 CEST 2018


I'm not sure how you are incorporating time period into your data structure. Typically we are looking at plots or assemblages as the rows and taxa as the columns. Time period adds a third dimension that could be added as blocks of rows. For example, depending on the resolution of your data, one approach would be to have up to 8 rows for each locality: Loc1.1000, Loc1.2000, . . . Loc1.8000. If a locality did not have certain millennia represented you would just leave them out.

David C

-----Original Message-----
From: Jessie Woodbridge <jessie.woodbridge using plymouth.ac.uk> 
Sent: Wednesday, April 18, 2018 3:42 AM
To: David L Carlson <dcarlson using tamu.edu>
Subject: RE: nMDS with R: missing values

Dear Prof Carlson,

Thank you for your reply. I'm using 'vegan' with 'vegdist' and 'bray'. I have a selection of datasets that cover different time periods (converted to z-scores), so a record that starts 5000 years ago would have missing data before this date when other records cover the last 8000 years. I need to build into the nMDS the fact that this isn't a zero score, but reflects absence of data. In the results it seems that datasets that cover the same time periods are being grouped together because they have zero values at the same times. I hope that makes sense. Any suggestions would be greatly appreciated.

There is no regular pattern to the position of the missing values, so I can't remove certain rows or columns. Thank you for the suggestion of using dist() and na.rm=TRUE. I will try using these approaches.

Kind regards,

Jessie

********************
Dr Jessie Woodbridge
Geography, Earth and Environmental Sciences B416, Portland Square, University of Plymouth, UK
01752 585920
Leverhulme Trust-funded ‘Changing the face of the Mediterranean’ project Leverhulme Trust-funded Deforesting Europe project

-----Original Message-----
From: David L Carlson <dcarlson using tamu.edu>
Sent: 17 April 2018 21:13
To: Jessie Woodbridge <jessie.woodbridge using plymouth.ac.uk>; r-help using r-project.org
Subject: RE: nMDS with R: missing values

I think you will have to provide some more information. What function/package are you using for nMDS (eg. isoMDS in MASS, monoMDS in vegan)? What function/package are you using to compute your distance/dissimilarities (eg. dist in stats, vegdist in vegan)?

Zero represents absence. It is not a missing value, so that part of your question is not clear.

The dist() function computes distance ignoring missing values if they are properly represented as NAs in the data. The vegdist() function does this if na.rm=TRUE.

The results will be affected by how much data is missing. It would be useful to know if the missing values are concentrated in particular rows or columns so that eliminating a few rows and columns could substantially reduce the percentage of missing values.

----------------------------------------
David L Carlson
Department of Anthropology
Texas A&M University
College Station, TX 77843-4352


0-----Original Message-----
From: R-help <r-help-bounces using r-project.org> On Behalf Of Jessie Woodbridge
Sent: Tuesday, April 17, 2018 2:05 PM
To: r-help using r-project.org
Subject: [R] nMDS with R: missing values

Dear All,

I was wondering whether anyone might be able to provide some advice with an nMDS / R problem. I’m trying to run nMDS on a dataset that contains many missing values and was wondering how I can account for the missing values when running nMDS? It seems as though the data are being grouped depending on where the zero values appear. Any suggestions greatly appreciated. Thank you very much in advance.

Apologies if this message isn’t relevant to you.

Kind regards,

Jessie

________________________________
[http://www.plymouth.ac.uk/images/email_footer.gif]<http://www.plymouth.ac.uk/worldclass>

This email and any files with it are confidential and intended solely for the use of the recipient to whom it is addressed. If you are not the intended recipient then copying, distribution or other use of the information contained is strictly prohibited and you should not rely on it. If you have received this email in error please let the sender know immediately and delete it from your system(s). Internet emails are not necessarily secure. While we take every care, Plymouth University accepts no responsibility for viruses and it is your responsibility to scan emails and their attachments. Plymouth University does not accept responsibility for any changes made after it was sent. Nothing in this email or its attachments constitutes an order for goods or services unless accompanied by an official order form.
______________________________________________
R-help using r-project.org mailing list -- To UNSUBSCRIBE and more, see 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.
________________________________
[http://www.plymouth.ac.uk/images/email_footer.gif]<http://www.plymouth.ac.uk/worldclass>

This email and any files with it are confidential and intended solely for the use of the recipient to whom it is addressed. If you are not the intended recipient then copying, distribution or other use of the information contained is strictly prohibited and you should not rely on it. If you have received this email in error please let the sender know immediately and delete it from your system(s). Internet emails are not necessarily secure. While we take every care, Plymouth University accepts no responsibility for viruses and it is your responsibility to scan emails and their attachments. Plymouth University does not accept responsibility for any changes made after it was sent. Nothing in this email or its attachments constitutes an order for goods or services unless accompanied by an official order form.


More information about the R-help mailing list