Attribute Reduction In Generalized Incomplete Interval-Value Information System

Rough set theory is a new tool for a data analysis proposed by the scholar Pawlak inPoland, classical rough sets theory is based on equivalence relations.Requirements are too strict,study of rough set extension model is extremely of important significance for the developmentof rough set theory.This article has majored in attributes reduction of generalized incomplete interval-valuedinformation system based on feature reations. We Define the feature relationship in generalizedincomplete interval-valued information system, generalized incomplete interval-valued objectiveinformation systems, and generalized incomplete interval -valued fuzzy objective informationsystems. We introduce upper-approximation ,lower-approximation and propeties. We give dis-tribution reduction and maximum distribution reduction, assignment reduction in generalizedincomplete interval-valued objective information systems based on feature relations, discuss therelationship between them, furthermore, judgement theorem and discernibility matrix of knowl-edge reduction are obtained, from which approaches to knowledge reductions can be provided ingeneralized incomplete interval-valued objective information systems based on feature relations.There are strong connections between rough set theory and Dempster-Shafer theory of evi-dence. It has been demonstrated that various belief structures are associated with various roughapproximation spaces . In this paper, we introduce the evidence theory in generalized incom-plete interval-valued information system, and discuss the approaches to attribute reductions ingeneralized incomplete interval-valued information system based on evidence theory.