likelihood ratio for censored data

#1
hi,

I'm working on multiply censored data. In order to check which is the law my data follow(weib vs exp), i need to do a test. For the moment, the best (and easier) test i find to test this, is the test of likelihood ratio...
but i can't find any kind of way to define the statistics of test...do you have ideas?

moreover, another option could be to use a kind of kolmogorov-smirnov test for censored data (an article of barr and davidson talks about this) but i can't find the table for a big number of data. Do you know if i can find this somewhere or if a monte carlo simulation if necessary to solve this problem?

thanks
 

vinux

Dark Knight
#2
" multiply censored data." I didn't get this part.
MINITAB has option for checking distribution of the survival data ( data with censors).
Otherwise, hope you have right censored data. In that case you can find the empircal distribution(for kaplan meir estimates, see this link http://faculty.vassar.edu/lowry/survival.html) and compare with the standard distribution and you can do Kolmogrov Smirnoff test.
Regards
VinuX