Declarative bias in Atris rule induction shell
The paper presents Atris rule induction shell in the light of
its declarative biases. Atris was developed to enable simple
use of different combinatorial optimisation
algorithms and noise-handling mechanisms, to enable simple
``cross validation'' experiments, experiments with different attributes
considered as class and different attribute subsets used for learning.
Atris currently incorporates seven local optimisation
algorithms and four heuristic estimates, enables use of simple
built-in background knowledge, use of pruning parameters and
validation of the induced hypothesis.