On Thu, 27 Feb 2003, Micky wrote:
I am applying WEKA's learning algorithms on a
set of data of string
type. I understand that the classifiers do not accept string attributes
and ID3 looks out for nominal values. What are nominal values? Can you
please give me some examples? Since I am not able to use string values,
what other types of values can I use?
There are two types of attributes: numeric (or continuous) and symbolic
(or nominal). String-type attributes are nominal attributes. Some
classifiers require numeric attributes, while others can handle either
type. If your data set is represented with string (nominal) attributes,
you should select a classifier such as J48 (decision tree). The
clustering methods, Cobweb and EM, can also handle nominal attributes.
See the Weka documentation to determine which classifiers handle which
kind of attributes.
Kiri L. Wagstaff, Ph.D. (kiri.wagstaff(a)jhuapl.edu) \|/
Science Applications Group, Space Department -O-
The Johns Hopkins University Applied Physics Lab /|\