JNCC2 is the Java implementation of the Naive Credal Classifier 2.   
   NCC2 constitutes an extension of Naive Bayes  towards imprecise   
   probabilities; it is designed to return robust classification, even on   
   small and/or incomplete data sets. A peculiar feature of NCC2 is that   
   it returns set-valued (or imprecise) classifications (i.e., more than   
   one class) when faced with doubtful instances. Imprecise   
   classifications are valuable as they clearly highlight doubtful   
   instances, preventing over-confident use of the issued judgments;   
   however, they still convey an informative content, dropping unlikely   
   classes. This could be appealing in contexts in which the   
   classification outcome is especially sensitive, as for instance in   
   medical area.   
      
   Extensive empirical investigation shows that NCC2 returns imprecise   
   judgments on instances whose classification is truly  doubtful; in   
   fact, Naive Bayes  achieves a much higher classification accuracy on   
   the instances precisely classified by NCC2, than on those imprecisely   
   classified by NCC2. We say that NCC2 isolates area of ignorance, i.e.   
   subsets of instances over which the accuracy of Naive Bayes sharply   
   drops.   
      
   JNCC2 is open source; it is released under the terms of the GNU GPL   
   license; it is hence freely available together with manual, sources   
   and javadoc documentation.   
      
   For more information, visit http://www.idsia.ch/~giorgio/jncc2.html.   
      
   [ comp.ai is moderated ... your article may take a while to appear. ]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)