Context-specific Sign-propagation in Qualitative Probabilistic Networks

Paper:

Context-specific Sign-propagation in Qualitative Probabilistic Networks

Appears:

Artificial Intelligence

Abstract:

Qualitative probabilistic networks are qualitative abstractions of probabilistic networks, summarising probabilistic influences by qualitative signs. As qualita tive networks model influences at the level of variables, knowledge about probabilistic influences that hold only for specific values cannot be expr essed. The results computed from a qualitative network, as a consequence, can be eaker than strictly necessary and may in fact be rather uninformative. We exte nd the basic formalism of qualitative probabilistic networks by providing for th e inclusion of context-specific information about influences and show that explo iting this information upon inference has the ability to forestall unnecessarily weak results.

Keywords:

Probabilistic reasoning, qualitative reasoning, context- specific independence, non-monotonicity.

Availability:

This paper is available as a compressed Postscript file.

Other information:

The postscript version is of the text accepted by the journal.