Research Of Rough Sets Reduction Based On Difference-identity View In Pansystems

Rough Sets theory, as a new math tool,is used to analyze and process the uncertainty, incompleteness and incompatibility data. Well, Pansystems theory which is a multilayerly network model interdisciplinary theory puts great emphasis upon mining the nature of something, it is a methodology that seeking the variation and connecting of something by the roots. By the research of Rough Sets based on difference-identity view in pansystem, we sum up the process of cognition that is the classifying process of knowing the difference-identity view of knowledge. The difference-identity view in Pansystems is a system concept, associated with set theory. On base of difference-identity view in Pansystems, this paper treats of reduction of knoeledge, classification of core and core-value in the decision table and reduction of knowledge spectrum, clears their essence of reduction.Firstly, the paper studys the essence of knowledge attribute reduction in Rough sets based on difference-identity view in Pansystems and draws a general conclusion, then gives primary conceptions which include the ablity of similarities and differences, and raises the algorithm of attribute reduction in repository based on difference-identity view.Then, the paper turns the general relationship into the complementary in Rough sets, and shows the classification of core and core-value by making of the definition with 3D- space, also founds out a method extracting all decision rules from the classification of core-value in knowledge spectrum reduction an gets the optimum decision rules through the weighting optimum principle.Finally, we put the methods into the aerial target recognition research of the land-based air defence systems in battleground nowadays. Based on the classifying about aerial target, Founding the system of evaluating index which comprises tactical characters and physical characters of target. On the basis of the knowledge reduction theory, a new algorithm for aerial target recognition is put forward. The method can search aerial target's available features, speed up the identification of aerial target and improve the recognition rate. The experimental results shows that the method is feasible and effective.