The Research Of Incomplete Grade Multi-granulation Rough Set

With the rapid development of network technology and communication technology,a variety of massive data has been generated.All fields are expected to obtain useful knowledge from massive,chaotic,noisy data.Rough set(rough set,RS)shows great advantages in the acquisition of uncertain,fuzzy knowledge.It can excavate the valuable information hidden in the data,without any other prior knowledge and additional information,just depending on the attributes of the data set.Multi-granulation rough set is a new extended model of rough set,which is used to approximate the concept of the object from multiple granularities.It has the obvious advantage in reducing the range of boundary region and improving the accuracy of the representation of the target concept.In real life,because of measurement error and other factors,there is some incomplete information but hidden rich knowledge of the data.In order to obtain more accurate knowledge in incomplete information system,this paper studies the incomplete multi-granulation rough set model and the granularity reduction method based on graded rough set.The main works of this paper are as follows:(1)Introduce the basic knowledge of classical rough set,and gives some examples to explain the basic concept of it.And then the development and research status of MGRS are introduced for complete and incomplete information system.(2)Introduce the extension model of single granularity and multi-granulation rough set based on tolerance relation,similarity relation and limited tolerance relation,and analyzes the advantages and disadvantages of each rough set model under different relations.(3)For the incomplete information system,the graded MGRS based on limited tolerance relation is proposed,which includes optimistic graded MGRS and pessimistic graded MGRS.Analysis the disadvantages of two models,proposes a variable graded MGRS model based on limited tolerance relation.The related properties and relations of the three models are studied.The superiority of variable graded MGRS is analyzed by examples and experiments.(4)Considering the weight of granularity,propose the incomplete weighted graded MGRS based on limited tolerance relation,and discuss its properties.Define granularity matrix,nuclear granularity and granularity importance formula for incomplete weighted graded MGRS.Propose a granularity reduction method.Based on the acquisition of nuclear granularity,the final granularity reduction set is obtainedby selecting granularity importance as heuristic information.