Dear Uaiers,
As David Heckerman said in his paper: Bayesian Networks for data mining,
Bayesian networks are usually employed in data mining learning tasks mainly
because they: (i) may deal with incomplete data sets in a direct way; (ii)
can learn causal relationships; (iii) may combine a priori knowledge with
patterns learned from data and (iv) help to avoid overfitting. However, the
knowledge acquired by such models are not so comprehensible for humans as
association rules are. Does this chacateristic represent an obstacle in
domains such as data mining where it is important to have symbolic rules or
other forms of knowledge structure?
I mean, may I say that in a data mining application, the bayesian methods
are appropriatted for the learning tasks (and so on), but for the data
visualisation they're not the most appropriated?
Thank you so much in advance for your attention,
Estevam.
- -_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-
Estevam Rafael Hruschka Junior
Curitiba PR.
Brasil
This archive was generated by hypermail 2b29 : Mon Jul 01 2002 - 11:03:07 PDT