Thursday, January 01, 2004


1. While running OxGauss to compute compensating variations, I used the time to also run Latent Class Model using Limdep. (Stata cannot do this. Their GLLAMM program is too complicated; yet, not specifically written for the kind of problem I am dealing with). But in Limdep, somehow I constantly got this error message saying "Divided by zero... " -- before even finished reading that popup message, the whole program crashed. This is frustrating. I remembered the same thing happened when I tried to run Random Parameter Logit less than a month ago. I contacted Bill Greene the author and sent him my code. He didn't think my code suggested any inherent problem. So it might have been the way I constructed the data, he probed. Fortunately I could finally do that using Gauss and OxGauss. Now, for the Latent Class Model, I again keep encountering the same problem in Limdep! Time is running and this program keeps crashing! I have tried two specs so far. Same results: crash.

2. Music: Sonny Rollins (Colossus). Too bad, I'm too burnt out I can't enjoy it very well.

3. Happy new year. Welcome 2004. 2003 has been nice but also very devastating. Somebody out there knows exactly what I mean.

No comments: