Title :
Reduction of rough approximation space based on matroid
Author :
Zhang, Shao-pu ; Wang, Xiao-Feng ; Feng, Tao ; Feng, Lin
Author_Institution :
Dept. of Math. & Phys., Shijiazhuang Tiedao Univ., Shijiazhuang, China
Abstract :
In this paper, we firstly review the definitions and matrix representations of a special pair of upper and lower approximation operators generated by a reflexive and transitive binary relation. Then the matroid generated by the reflexive and transitive binary relation of rough approximation space and its properties are discussed. Simultaneously, we consider how to generate a rough approximation space by a matroid (satisfying ∀X ⊆ U, ∀y ϵ X, if x ∉ d(y) or y ∉ d(x), then |d(X∪{x})| = |d(X)|+|d(x)|), and we get that the special upper approximation operator equals to the closure operator in the matroid. Moreover, we give the axiomatization characterization of the rough approximation operators. Finally, we give an algorithm to reduce the information system using the definitions of matroid.
Keywords :
approximation theory; combinatorial mathematics; matrix algebra; axiomatization characterization; information system; lower approximation operators; matrix representations; matroid; reflexive binary relation; rough approximation operators; rough approximation space reduction; transitive binary relation; upper approximation operators; Bismuth; Legged locomotion; Approximation operators; Closure operator; Matroid; Reduction; Reflective and transitive binary relation;
Conference_Titel :
Machine Learning and Cybernetics (ICMLC), 2011 International Conference on
Conference_Location :
Guilin
Print_ISBN :
978-1-4577-0305-8
DOI :
10.1109/ICMLC.2011.6016728