Research On Algorithms For Updating Approximation In Multigranulation Rough Sets

With the advent of the information technology revolution,data sets containing information have become large and complex,sometimes difficult to use.Dynamic computing is an effective way to solve these problems.Since the multi-granularity rough set was proposed by Qian Yuhua et al.,in order to expand the application range of multi-granular rough sets,many extensions have been proposed by the authors,including neighborhood multi-granulation rough sets,many Granularity variable precision rough set,multigranularity incomplete rough set,multi-granularity variable precision decision rough set.Since data sets are being huge and complex,it is more and more difficult to compute the approximations of multi-granulation rough sets.This paper attempts to solve the problems by dynamic approaches.This paper mainly involves the following aspects:(1)For the multi-granulation rough set,first,this paper discusses that some elements in the universe are of no need to be discussed because they have either been determined to belong to each of the approximations,or have been determined not to belong to any of the approximations.Based on this fact,algorithms based on vector matrix is given.Finally,several experiments are conducted to verify the effectiveness and validity of the algorithm.(2)For the neighborhood multi-granulation rough set,first,this paper discusses that some elements in the universe are of no need to be discussed because they have either been determined to belong to each of the approximations,or have been determined not to belong to any of the approximations.Based on this fact,algorithms based on vector matrix is given.Finally,several experiments are conducted to verify the effectiveness and validity of the algorithm.(3)Finally,this paper applies the fact that some elements in the universe do not need to be discussed whether they belong to the approximations to improve the approaches that based on the relation matrix,and proposes the corresponding algorithms for updating approximations in neighborhood multi-granulation rough sets.