Skip to content

NA's in segmented

3 messages · Mélina Cointe, Ivan Krylov, PIKAL Petr

Mélina Cointe
# 
Hi,

I?m contacting you because I have some trouble with the slope() function of segmented. When I plot the predicted value everything seems to have worked well. But when I use slope(), the slope for the first segment is a line with 0 and Nas? Also I can get a negative slope whereas on the graph it?s clearly a positive slope?

Here are the lines I?m using :

Mod = lm(as.numeric(data[,"meanMSD25"])~(as.numeric(data[,"Slice"])), weights=1*sqrt(as.numeric(data[,"Detec"])))
  x = (as.numeric(data[,"Slice"]))
  o <- segmented(Mod, seg.Z=~x, psi=NA, control=seg.control(display=FALSE, K=2))

Thanks in advance for your help,

Best regards,

M?lina COINTE
1 day later
Ivan Krylov
# 
Hi M?lina,

If you don't get an answer here, consider running maintainer(segmented)
and contacting the e-mail address shown by the command.

On Tue, 8 Mar 2022 20:44:35 +0000
M?lina Cointe <melina.cointe at outlook.fr> wrote:

            

      
      
      
    

        
It would help if you show us a small sample of your data. Use dput() to
produce a plain text representation of it. Otherwise it's next to
impossible to reproduce the problems you're having.

            

      
      
      
    

        
When you post in HTML, the list removes the HTML version for safety
reasons, and all we're left with is the plain text version
automatically prepared by your mailer. This plain text version can be
severely mangled and hard to understand. Please post in plain text:
http://www.R-project.org/posting-guide.html

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      PIKAL Petr
    

    

      
      #
    

  


  

      
Hi

1. Do not use HTML formated mail, your message is scrambled
2. With your code only you can get result all of us get error

            

      
      
      
    

        
Error in data[, "meanMSD25"] : 
  object of type 'closure' is not subsettable
Try to provide reproducible example.

3. with lm it is not necessary to use data[,"meanMSD25"], you should provide your data object to data option in lm call (it is silly to call your data data - fortune("dog") applies here)
4. why do you use as.numeric(data[,"meanMSD25"]), shouldn't the column be numeric itself?

So my  wild guess is that you read the data into R wrong way and instead of numeric they are character, which could be one source of your problem.

Cheers
Petr