Segmenting data and Segclust2d

[Noah Osterhoudt]

Hi Everyone,

I am attempting to analyze multiple black vulture’s home ranges collected from transmitters over a 12 month period. Many of my birds variograms do not reach an asymptote with the standard variogram function of the entire dataset, thus I am looking to segment the data. Looking at their movements plotted, it seems clear there are home ranges being formed for periods of time 3-4 months, with a long dispersal movement happening in between. The overall goal is to segment the data to periods when a home range is established, and to use mean() to combine these into one larger range for the whole year.

 

I was wondering if anyone would be able to provide help or advice with using the segclust2d package that has been mentioned previously in the group? I’ve been playing around with this, and have yet to find results I felt confident in. Namely I have been trying to use the segmentation() function to segment the data automatically into areas where a home range may have been established. Although when I do this, it feels as if it’s just segmenting as much as possible and leaving segments that span large amounts of space, instead of segmenting the dispersal movement into it’s own segment phase. If anyone has example code or data they’d be able to share that would be great. Also, just any general advice or directions of where to look would be appreciated as well.

 

Thank you!

 

-Noah

[Kerry Kilshaw]

Hi Noah

I've been playing around with this for the past few months, I'll pm you next week if that suits? It works pretty well once you get some of the variables sussed out (I think anyway!)

Best

Kerry

[Sarah P]

Hello, I am out in the field from March 10-March 14. I will be back Monday March 17.

Best,
Sarah

> --
> You received this message because you are subscribed to the Google Groups "ctmm R user group" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to ctmm-user+...@googlegroups.com.
> To view this discussion visit https://groups.google.com/d/msgid/ctmm-user/3fffa9ce-7042-42e6-b294-2f566285b3c7n%40googlegroups.com.

[Noah Osterhoudt]

Kerry, 

That would be much appreciated, thank you!!

Best,

[Jake Straub]

-Kerry Would also mind sharing your  summary of your expertise with Segclust2D on this public forum?  I think many (including me) would like to learn what you have to say.  

My students and I started using Seg2clust with the hopes it would be an objective approach to when (and when not) to segment a dataset after Variograms indicated it might need it.  But so far, we have been underwhelmed with this approach.  In our experience the Seg2cluster algorithms tend to 'oversegment' quite consistently. 

Thanks

Jake

--

[Kerry Kilshaw]

Hi Jake

I certainly wouldn't describe it as expertise, more like bumbling along until something worked! But I can certainly copy you into the email with Noah and a couple of others, it would be great to see if the workaround I have tried works for others and then we could make some comments on the group page here...?

Hopefully I'll be able to send the email tomorrow or the next day, hope that's ok Noah.

Best

Kerry

[Kerry Kilshaw]

Hi Noah

Could you send me an email to contact you on, I couldn't see it on this chain.

Thanks

Kerry

[yu lei]

Hi Kerry

    
I have a similar question as Noah. I guess that adding variables like step length or net displacement might work. Would you be willing to share your ideas and code? We can continue the discussion here or via email at leial...@126.com.

Thanks

Yu

[Kerry Kilshaw]

Hi Yu

So sorry only just seen your message, will email you unless you already have it sorted? Not sure if I'll be much help though!

Best

Kerry