Skip to content

Inverse Gaussian Distribution

7 messages · David Winsemius, Dennis Murphy, Louisa

Original
Louisa
# 
Dear, 

I want to fit an inverse gaussion distribution to a data set. 

The predictor variables are gender, area and agecategory.
For each of these variables I've defined a baseline 

e.g. 
#agecat: baseline is 3 
data<-transform(data, agecat=C(factor(agecat,ordered=TRUE),
contr.treatment(n=6,base=3)))

The variable 'area' goes from A to F (6 areas: A,B,C,D,E,F)

How can i manipulate the data to set the baseline of area to C? 
R is producing errors when I'm trying to do so. 

I'll be very thankful for any help you can provide. 

Louisa
David Winsemius
#
On Jan 3, 2011, at 3:03 PM, Louisa wrote:

            

      
      
      
    

        
In all likelihood it's a factor. Try
area <- factor(area, levels=c("C", "A","B","D","E","F") )


If not, then you need to provide more information. Read the Posting  
Guide.

            

      
      
      
    

        
David Winsemius, MD
West Hartford, CT

  
  


  


      

    

    

      
      

        
      

    

    

      
      

        


  

        

      Louisa
    

    

      
      #
    

  


  

      
Thank you!

But i'm wondering: 

if you run
 
area <- factor(area, levels=c("C", "A","B","D","E","F") ) 

then you are transforming only 'area',  aren't you? 

isn't it possible to transform the whole data like i did for agecat 
but now for area and area C as baseline, 

or are you doing so when you run

            

      
      
      
    

        
and then run the model with area as predictorvariable:

 > model <- glm(Y~ agecat+gender+area,...)

My question is if i can run it as follows and still have a right solution :

            

      
      
      
    

        
I'll be very grateful for any help you can provide!

Kind regards, 

Louisa

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      David Winsemius
    

    

      
      #
    

  


  

      
On Jan 4, 2011, at 7:50 AM, Louisa wrote:

        

            

      
      
      
    

        
Of course, but may I point out that you did no earlier say that you  
may have attached a dataframe containing "area". There are many  
confusing situations that arise when using attached dataframes and  
this may be one of them. Many experienced R users avoid attaching data  
objects like the plague.

            

      
      
      
    

        
Is there any reason that you suspect this will not work? Why not try  
it ... now assumimg that you failed to tell us that area was a column  
in a dataframe that is. (One set of parentheses appear extraneous.)

data <-transform(data, area = factor(area, levels=c("C",  
"A","B","D","E","F") )

(Untested in the absence of reproducible example ,,, but it looks as  
though it might work.)

            

      
      
      
    

        

      
      
    

        
David Winsemius, MD
West Hartford, CT

  
  


  


      

    

    

      
      

        


  

        

      Louisa
    

    

      
      #
    

  


  

      
Thank you again David! 

I did not try it yet, cause neither the dataset nor R is on this computer.
I'll try it in a few hours, as soon as possible, when I'm on my personal
computer. 

I'll let you know if it works. I'm really curious!

Thank you for your time!

Best Wishes, 
Louisa