[R] Incomplete output with `sn' library package

[Uwe Ligges]

The derived information matrix is not of full rank for your data.
See the code of the functions, which is not that hard to read.

Uwe Ligges





MANASI VYDYANATH wrote:
> You have my sincere apologies for the incompleteness of my message.   
> I have given the details below, including my dataset and my code.
> 
> I'm using R, version 2.5.0. My OS is a Mac, (version Tiger).
> 
> The sn package is Version 0.4-1
> 
> My code was as follows:
> 
>  > mydata <- read.table(url("http://www.statsci.org/data/oz/ 
> ais.txt"), header = T)
>  > attach(mydata)
>  > a <- msn.fit(X = cbind(1,Ht,Wt), y = BMI, control = list(x.tol=1e-6))
>  > b <- msn.mle(X=cbind(1,Ht,Wt), y=SSF)
>  > a
>  > b
> 
> My problem is that neither the "a" nor the "b" output gives me any  
> standard errors - those should appear under <$se>. In both the  
> regressions, this field is left blank with "NA" under it. I would  
> appreciate some help on this matter - are the standard errors not  
> supposed to appear here, or is there something else I should put into  
> the inputs?
> 
> Thank you once again for your time,
> 
> Manasi
> 
> 
> 
> On Aug 30, 2007, at 5:47 AM, Uwe Ligges wrote:
> 
>>
>> MANASI VYDYANATH wrote:
>>> Dear R users:
>>> I have a question regarding the output for two of the functions  
>>> in  the `sn' package, which deals with the mle fitting of skew  
>>> normal  curves to linear regressions. I'm using the examples and  
>>> the dataset  given as an example in the online documentation for  
>>> this package, for  the functions `msn.fit' and `msn.mle'. I'm  
>>> following the example code  in the documentation for these two  
>>> functions exactly.
>>> Part of the data output is supposed to be "se", which gives the   
>>> standard errors of the estimated coefficients. This particular  
>>> value  comes out as being "NA" in the examples given, but there  
>>> are three  coefficients in each case and no numerical problems  
>>> about why the  standard errors cannot be calculated.
>>> Am I setting this program up right? Is there some other command I   
>>> should use (or an option I need to use) to get the output to  
>>> display  standard errors of the coefficients?
>> We cannot know if you use it right, since you have not given any  
>> details on
>>
>> OS, R version, sn version, and particularly a reproducible example.
>>
>> As each R-help message tells in the footer:
>>
>> "PLEASE do read the posting guide http://www.R-project.org/posting- 
>> guide.html
>> and provide commented, minimal, self-contained, reproducible code."
>>
>> Uwe Ligges
>>
>>
>>
>>
>>
>>
>>
>>> Thank you for your time in reading this question -
>>> Cordially,
>>> Manasi
>>> ______________________________________________
>>> R-help at stat.math.ethz.ch mailing list
>>> 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.
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> 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.