On Mon, 13 Jul 1998 17:06:59 +0100, Ian Eiloart wrote:
>I take the point, though. If you can codify the relevant diagnostic
>criteria, and get an adequate sample of case histories, then a computer
>can be a useful diagnostic tool. But this is actually just a case of
>recalling information reliably - although the interface for generating the
>recall has some rules in it that relate symptoms to causes.
News item on R4 today mentions St George's doing just that with brain
tumour MRI to avoid un-necessary exploratory operations.
However, for the work I do, a big part of the decision making process
involves intuition, 'feel' and prior knowledge of the patient in a
non-linear way.
These non-linear processes of the human brain are difficlt to emulate
with computers. There are various experiments with neural networks
though.
Ahmad
________________________________________
Dr Ahmad Risk
http://mednetics.org
home: +44 1273 748198
work: +44 1737 240022
fax: +44 1737 244660
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|