Hi,

I wonder if anyone can help me. I want to do some research into machine
learning, but lack access to any academic library. My workplace can
order any papers from any journals on my behalf, but the trick is
knowing which ones.

I'm _not_ interested in the actual development of the algorithms and the
underlying maths. My area of research is into the linguistic properties
of different types of text in its relation to machine learning and auto
classification. In particular I want to research into the nature of
medical content, vs standard (control) content like the Reuters news
set. The relevant areas would be around feature weighting, tuning
algoirthms, preparation of text (e.g. creating stopword lists), and so
on. I also wish to research on cross domain training classification -
i.e. the effects of training a categoriser on text written for doctors
and then using it on text written for nurses or consultants.

So, as you see I'm more interested in the nature of the text and the
selection of training text, than the specifics of different algorithms.

I have some ability to search through abstracts online, but find little
that's directly relevant. If anyone has any pointers to journals or
papers that might be good starting points, that's would be very useful.

Jon
-- 
  Jonathan Peterson
  jon@snowdrift.org