Saturday, July 12, 2008

As my gibbs sampling wanders over yonder distribution

So I'm running my final run of simulations, results obtained so far for the first of three specifications are theoretically incorrect for 3 out of 6 coefficients. Considering that the whole point of what I'm doing is to ensure that signs are correct, this puts things in a not happy place. Ah well, I tried for days to figure out what was going wrong. Now is time to write, not to find out why.

I don't usually do these look-at-my-life posts but considering I'm running 40,000 iterations I've tons of time.

I like Lemmy's voice; he cannot sing of course, but its in tune. Like Brian Johnson of AC/DC. It's an admirable talent - to not be able to sing but stay in tune.

Maybe my constraints aren't right, but I've taken them from a published paper in Journal of Econometrics, could they be wrong? Unlikely.

Welcome home, it's been too long we've missed you. Feel our bodies breathing as you try to stop believing.

So its a Saturday and the only 2 other human beings in the department is a Moroccan and an Italian. Italian dude is asleep in office, he frequently stays overnight. There is also at least one moth, two cockroaches I think that's it. Yes there are cockroaches.

Maybe I should get a coffee from somewhere.

Charlotte the harlot, who made you this way, to make your rent as a bloody whore?

Read Barkha Dutts article in HT on what's going on in Stephens. Its a bit saddening. But, in a way, it's good. Sometimes there is a tendency to over romanticize college. And the things that are happening now, they serve to highlight the stuff that isn't quite right. So this may provoke some reaction that could have a positive effect. It's ugly to have this stuff on newspapers, but better it to be in newspapers than hidden away.

3000 iterations left to go!

2 comments:

Unknown said...

oi when do you get to india?

k said...

in 2 weeks time.