This package was designed to manipulate data with DRSA (Dominance-Based Rough Set Approach) usage. Still in development.
Implemented the following algorithms:
- DomLEM algorithm (DRSA),
-
VC-DomLEM algorithm with
$\epsilon$ -consistency (VC-DRSA) - extended VC-DomLEM algorithm (VC-DRSA with Naive Bayes)
Supported two classification methods - "standard", based on the intersection of supported classes, and "advanced", based on the score (article)
Class approximations can be generated with thresholds from this article:
-
$\mu$ (gain type measure) -
$\epsilon^*$ (cost type measure)