topic Re: Estimating parameters of different models in Statistical Procedures

Estimating parameters of different models

Camp — Sun, 03 Jul 2016 07:14:44 GMT

Hi Everyone,

I have just stuck up on a problem – estimating parameters (scale and shape) to fit different survival models like Exponential, Weibull, Log-normal, Log-logistic and Gompertz (given Kaplan Mirror survival data).

For this I have attached one excel file having two sheets – DATA (to be used by you as raw data) and OBJECTIVE (to understand the problem) and my aim is to estimate the constants – scale & shape parameters, variance & covariance matrix and AIC and BIC values that you can see in OBJECTIVE sheet.

Also OBJECTIVE sheet has some green highlighted columns (C to G) that you can ignore. Once we estimate the point estimates these distributions can be fitted (column C to G) so just ignore these columns. However, these columns might help to understand the problem better.

So, I am interested in estimating these constants and variance-covariance matrix and AIC and BIC statistics for each of five statistical distributions. Please provide me the required SAS codes for the same.

Thanks you!

Re: Estimating parameters of different models

PGStats — Sat, 02 Jul 2016 19:29:28 GMT

What is the "SAS code" that you tried?

Re: Estimating parameters of different models

Reeza — Sun, 03 Jul 2016 02:36:17 GMT

Do you have the raw survival data?

Re: Estimating parameters of different models

Camp — Sun, 03 Jul 2016 06:17:29 GMT

Hi Reeza,

Please find the attached file "Fitting" having new sheet RAW_DATA - for raw survival data.

Thank you!

Re: Estimating parameters of different models

Reeza — Sun, 03 Jul 2016 06:21:49 GMT

For fitting curves look at proc univariate.