i am using PROC LIFEREG for a censored data to estimate the paramters for a weibull distribution, however i want to compare the estimated parameters by using Least Square and Max likelihood method.

I have read the help file, there is no option in LIFEREG to choose LS or MLE.

anyone know how to choose the calculation method for this in SAS?

thanks, i appreciate.

Parameter estimation in LIFEREG using MLE only. Least square method cannot be used because of censored data.

Parameter estimation in LIFEREG using MLE only. Least square method cannot be used because of censored data.

for censored data, for small samples size data, LS sometimes have a better estimation than MLE, that is the reason why I want to compare MLE and LS.

What about I extract failure data from the censored data, then using LE to estimate?

for censored data, for small samples size data, LS sometimes have a better estimation than MLE, that is the reason why I want to compare MLE and LS.

What about I extract failure data from the censored data, then using LE to estimate?

I guess, you will take the maximum available information for censored observations. You can do that. But it will be a biased estimation ( under estimation).

And when we compare LS and MLE, many times LS shows a close fit when we compare with response (say Y) and pred(Y).

Because in LIFEREG, Y is not the response variable most of the times. Response will be Log(Y) or Logit(Y) or... depending on the distribution.

See this for more info
http://math.uc.edu/~scl/computing/lifereg.pdf