Skip to content
Prev 5891 / 7420 Next

[EXTERNAL] Multiply censored observations

Hi Brian, 

Thank you very much for your response.  I haven't tried the approach you suggested yet, but it seems it does offer a great deal of flexibility.  

So far, I looked into censored (MLE) regression as implemented via the function cenreg in the NADA package.  

Do you have any examples of R code that you would be able to point me to for the quantreg() approach? Or know of any articles where this approach was used?  
If yes, I would love to read up more on it.  

I always get frustrated when I have to apply statistical methods originally developed in the context of assessing trends in water quality parameters.  Their implementation in R 
seems to be patchy and often reduced to cases where either no censoring is present or there is a single censoring limit.  If we start throwing in things like seasonality, ties 
within a season, temporal/seasonal correlation, data missigness, etc., things get even more frustrating.  I would imagine quantreg() itself may struggle with correlation issues, for 
example.  There are some useful USGS packages out there, but many of them seem to be "orphaned" so one feels a bit unsure about how reliable they are if they are no longer 
maintained.   
  
Thank you again, 

Isabella 

Isabella R. Ghement, Ph.D.
Ghement Statistical Consulting Company Ltd.
301-7031 Blundell Road, Richmond, B.C., Canada, V6Y 1J5
Tel: 604-767-1250
E-mail: isabella at ghement.ca
Web: www.ghement.ca



On Mon 28/01/19  2:50 PM , "Cade, Brian" cadeb at usgs.gov sent: