fuzzy-c-mean (FCM), questions.

MCC (mcc@access.ch)
Tue, 25 Jun 1996 14:28:50 +0200


My problem is:
1) - is it allowed to cluster different types of variables with a FCM,
for example: bloodpressure, pulse and temperature from
different patients ?
if not: what kind of possibilities ar recomanded to cluster the
patients data.
2) - what shall I do with missing values ?
3) - how many patterns do I need to cluster them with FCM, if each
pattern has 100 dimensions ?.
4) - has the position of a variable in the dimension array on
influence of the result of the FCM ?
5) - are there other fuzzy cluster methods to classify diseases ?
Thank you
H.J. Casal