enismirdal: (heart)
[personal profile] enismirdal
Woooh. I'm having one of those so-tired-I'm-dizzy evenings. This being in spite of having got 9 hours' sleep last night.

I think it's London. And the Tube. It's been a busy day.

I got into the lab assuming that at 11am I'd leave again and get the Tube to Kew for a meeting, only to have a semi-frantic back and forth e-mail with 2 supervisors about timing and so on. The meeting was moved back 1 hour, leaving me with an extra hour in the lab to motivate myself and do something useful. So I messed about with some more code and read a few papers.

Tube to Kew. Got lunch from the organic place in Kew Village, made mistake of not reading the ingredients in enough detail and bought a wrap with 3 different ingredients I don't really like. Oops. Oh well. Realised late October is really rather cold, and shivered.

Meeting was encouraging and made me pretty optimistic (as in, the likelihood of lots of papers coming out of the project), and as usual filled with ideas. We got going again after an hour, hopping back on the good ol' District line and heading to South Ken and Imperial college.

Another meeting, this time with a statistician at Imperial. Not much was decided, but enthusiasm was shared. It seems the data I have to analyse is annoyingly resistant to being analysed! The statistician seemed to imply that there weren't any programs with the compuational power to tackle the enormous unwieldy function required to test the significance of our data. It left me wanting to know more, as it sounds like the best way to do it would just be to write the function in a pretty hardcore programming language, press "start" and then come back 3 days later to see the results. In any case, running statistics on the data will be a fairly key part of actually getting my papers, so it must be done. Even though it sounds Scary.

Once more on to the Tube. Delays on District Line, signal failure at Monument. Train crawling along. So changed to Victoria Line and thence to Central Line, which left me with a 10 minute walk home but was so desperate to get out of the obnoxious Tube system that I didn't care any more.

One of my Amazon orders arrived. Talk about overpackaging!! The box is around 35x25x15cm. And contains...
...one 1.5m patch cable. Coiled up. This would have fitted in A5 JIFFY BAG!!

Way to go with conserving world resources, Amazon!

Date: 31 Oct 2006 20:21 (UTC)
cjwatson: (Default)
From: [personal profile] cjwatson
Without knowing exactly what the statistician in question is talking about, it sounds like he/she might mean that the significance-testing function is intractable, i.e. its run-time increases exponentially with the size of the problem; in which case it won't make much difference how enormous a computer or how efficient a programming language you use, as the benefits would be swallowed up very quickly as the number of data points grows. With such problems the best you can do is often to figure out a good approximation.

I haven't heard of any such problems in statistics before, though, so I could be completely off-base.

Date: 31 Oct 2006 20:25 (UTC)
From: [identity profile] enismirdal.livejournal.com
Aah, I suppose that's possible. *hides from size of maths-monsters* Hopefully I'll learn more when I next meet up with her and a computer-specialist colleague to discuss things further. :) She's sending us a thesis written by one of her students to give us some background, which should make good reading as apparently the student was Not Very Good so a lot of the results must be taken with a pinch of salt!

Date: 31 Oct 2006 20:22 (UTC)
From: [identity profile] daeomae.livejournal.com
Goodness!! You poor dear!! *hugs* Get some rest!


*Mwah*

Date: 31 Oct 2006 20:26 (UTC)
From: [identity profile] enismirdal.livejournal.com
Hee! Life seems quite determined that I don't deserve rest at the moment!

Date: 31 Oct 2006 20:42 (UTC)
emperor: (Default)
From: [personal profile] emperor
R. R is the way to go for hard-core stats.

Date: 31 Oct 2006 22:36 (UTC)
From: (Anonymous)
Surely R 'n' R is what she needs but isn't getting?
Abner
R 'n' R = rest and recuperation!

Date: 31 Oct 2006 23:16 (UTC)
From: [identity profile] ilanin.livejournal.com
No, Rock and Roll. Rest and Recuperation is R&R.

Date: 1 Nov 2006 07:12 (UTC)
From: (Anonymous)
Yes it is isn't it. My brain isn't quiet right.
Abner

Date: 31 Oct 2006 22:37 (UTC)
From: (Anonymous)
Walking beats tube most days. Indeed I can do Barbican to Waterloo in only 5 minutes more than the tube takes!
Abner

Profile

enismirdal: (Default)
enismirdal

October 2025

M T W T F S S
  123 45
6789101112
13141516171819
20212223242526
2728293031  

Style Credit

Expand Cut Tags

No cut tags
Page generated 29 January 2026 22:41
Powered by Dreamwidth Studios