Optimal Scale Selection And Optimizing In Multi-scale Decision System Under The Environment Of Group-objects Updating

As a kind of important rational database,multi-scale decision system has attracted much attention and become a research hotspot.At the same time,data may be update dynamically over time in the practical application,so how to get useful knowledge and effective data mining method through the discovery of the optimal scale and optimizing rule induction from the massive data of group-object updating is one of the important tasks of the research of the multi-scale decision system.Based on the granular computing and rough set theory,this thesis first defines three kinds of consistencies in the multi-scale decision system(i.e.complete consistency,partial consistency and inconsistency),as well as their optimal scale and rule induction methods.And then,this thesis discusses the changes of optimal scale and rule induction under group-objects updating.Finally,the feasibility of the proposed method is verified by conducting some datasets.The main contributions are as follows:(1)Three kinds of consistencies of the multi-scale decision system are presented.(2)The optimal scales of the multi-scale decision system and objects are given,and their relationships are discussed.(3)The changes of the optimal scale and rule induction in the multi-scale decision system are investigated under the environment of the group-objects updating.