Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: AW: st: Weibull


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: AW: st: Weibull
Date   Mon, 9 Mar 2009 12:47:11 -0000

The archive shows the contributions to the thread. 

You can get Weibull deviates in at least three ways:

1.  by using -generate- directly, through something like 

gen weibull = b * (-ln(runiform()))^(1 / c)

2. by using Bobby Gutierrez's more general -gengammareg-

3. by using Joe Hilbe's -rnd-. 

As I see it, 3 offers no advantage over 1 beyond the fact that you don't need to know the formula with 3. 2 has a major advantage over 1 or 3 in that it can cope with (e.g.) generating deviates conditionally on linear predictors. 

As all this is evident by looking at help files, I am not sure what scope remains for a detailed discussion. If people don't understand, they should ask specific questions. 

Also, be aware that -rnd- is a not an equivalent to -simulate-. 

Nick 
n.j.cox@durham.ac.uk 

Muhammad Riaz

Well, I will stop adding more on this topic as it is going to be an issue now.
my aim was just to help, I don't remember who else has provided another method to generate data from weibull. It will help Nikolao if somebody could tell him how to use -rnd- instead -simulate- in his problem. Also comparison of  the two methods will give better understanding of the topic if somebody else has thrown another method. 
I welcome and appreciate your comments.
 

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index