he he, you know what they said about curiosity.
I do a lot of data mining, well it more text mining actually. I’m working on a project to extract cytogentic mutation details from within known cancers, so far mainly leukaemia’s, from the abstracts and full texts of published papers. I pull papers down from pubmed (an online databank of most medical papers since the early 60’s).
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi
and send millions of reg exp's across them looking for word patterns and matchs. It pulls out the syntax that relates to the cytogenetic breakpoints that can be seen in some cancers. This is dumped in a db, which can later be used (hopefully) to aid research. I really should get around to designing a web front for the db, but that not really my goal just yet.
The project was running as part of a group at the Sanger Institute Cambridge,
http://www.sanger.ac.uk
But I left there quite a while ago. I maintain the project in my personnel time, of which I get precious little.
Anyway the mining is done using perl reg exp’s and at the volume we’re talk about it become hugely processor hungry stuff. I don’t have the luxury of the sangers farm, or their new blades anymore so I have to do it from home.
Hope that clears things up for you.
Oh, and I should be picking up a 6800gt soon, so there goes more of my free time.
