This paper introduces the Dezert-Smarandache (DSm) Field and Linear Algebra of
Refined Labels (FLARL) useful for dealing accurately
with qualitative information expressed in terms of
qualitative belief functions. This work extends substantially our previous works done in DSmT framework
which were mainly based on approximate qualitative
operators. Here, new well justified accurate basic
operators on qualitative labels (addition, subtraction,
multiplication, division, root, power, etc) are presented.
The end of this paper is devoted to an exemples of
qualitative fusion rules based on this new FLARL
approach for decision-making support.
Category: Artificial Intelligence