Semantics of Rough Sets From theory to applications (through - - PowerPoint PPT Presentation

Semantics of Rough Sets

 From theory to applications (through

semantics understanding).

 Motivated by Dubois, D. and Prade, P.

The three semantics of fuzzy set, Fuzzy Sets and Systems, 90, 141-150, 1997.

Rough sets and concepts

 Rough sets for modeling concepts, or

more precisely the extensions of concepts.

Concepts

 A concept, in the classical view, is

defined by a pair of intension and extension.

 The extension: instances to which the

concept applies.

 The intension: a set of singly necessary

and jointly sufficient conditions that describe the instances of the concept.

Concepts: idealization

 Concepts are assumed to have well-

defined boundaries and their extensions can be precisely defined by sets of objects.

Concepts: reality

 Concepts in many practical situations

cannot be precisely defined and their extensions may not be given exactly by sets of objects.

Concepts: generalization

 Concepts with grey/gradual boundaries.  Partially known concepts.  Undefinable concepts.  Approximate concepts.  … … … …

* Last three situations were discussed by Marek, V.W.

and Truszczynski, M. Contributions to the theory of rough sets, Fundamenta Informaticae, 39(4) , 389 - 409, 1999.

Concepts with grey/gradual boundaries

 There is not a well-defined boundary that

differentiates the instances from the non- instances of the concept. For some objects, the concept only applies partially instead of fully.

 Fuzzy sets.

* Zadeh, L.A. Fuzzy sets, Information and Control, 8, 338–353, 1965. * Dubois, D. and Prade, P. Fuzzy rough sets and rough fuzzy sets, International Journal of General Systems, 17, 191-209.

Partially known concepts

 An object may actually be either an

instance or not an instance of a concept.

 Due to a lack of information and

knowledge, one can only express the state of instance and non-instance for some objects, instead of all objects.

Partially known concepts

 One has a partially known concept defined by a

lower bound and upper bound of its extension.  Interval sets



Y.Y. Yao, Interval-set algebra for qualitative knowledge representation, Proceedings of the Fifth International Conference on Computing and Information, 370-374, 1993.

Undefinable concepts

 In general, the intension of a concept

may be defined by using a logical language, such as the decision logic language in rough set theory and knowledge representation languages of description logic.

Undefinable concepts

 It may happen that some objects

cannot be differentiated due to the use

• f a fixed and limited set of attributes

for their description. The language may not be able to define certain sets

• f objects that are the extensions of

some concepts. That is, we have some undefinable concepts with respect to a particular language..

Undefinable concepts

 We have to approximate undefinable

concepts by definable concepts.

Rough sets.

 Pawlak, Z. Rough sets, International Journal of

Computer and Information Sciences, 11, 341-356, 1982.

Approximate concepts

 One knows the exact extension of a

concept and the language can precisely define the concept.

 The description may be too complex to

be of any practical value; it may be difficult to understand and manipulate.

Approximate concepts

 This may require us to approximate the

concept by some concepts with simple descriptions.

 Rough sets.

* Marek, V.W. and Truszczynski, M. Contributions to the theory of rough sets, Fundamenta Informaticae, 39(4) , 389 - 409, 1999.

Thank You!

http://www.cs.uregina.ca/~yyao/rough_set.html