ORIGINAL ARTICLE
On multigranulation rough sets in incomplete information system
Xibei Yang
•
Xiaoning Song
•
Zehua Chen
•
Jingyu Yang
Received: 20 April 2011 / Accepted: 7 October 2011 / Published online: 20 October 2011
Ó Springer-Verlag 2011
Abstract Multigranulation rough set is a new and inter-
esting topic in the theory of rough set. In this paper, the
multigranulation rough sets approach is introduced into
the incomplete information system. The tolerance relation,
the similarity relation and the limited tolerance relations
are employed to construct the optimistic and the pessi-
mistic multigranulation rough sets, respectively. Not only
the properties about these multigranulation rough sets are
discussed, but also the relationships among these multi-
granulation rough sets models are explored. It is shown that
by the multigranulation rough sets theory, the limited tol-
erance relations based multigranulation lower approxima-
tions fall between the tolerance and the similarity relations
based multigranulation lower approximations, the limited
tolerance relations based multigranulation upper approxi-
mations fall between the similarity and the tolerance rela-
tions based multigranulation upper approximations. Such
results are consistent to those in single-granulation based
rough sets models.
Keywords Incomplete information system Limited
tolerance relation Multigranulation rough set
Similarity relation Tolerance relation
1 Introduction
Rough set theory [9–12], proposed by Pawlak, has been
demonstrated to be a well-established mechanism for
uncertainty management in a wide variety of applications
related to artificial intelligence. It not only provides a novel
paradigm to deal with uncertainty, but also has been suc-
cessfully applied to feature selection, machine learning
[2, 5], rule extraction [26], data mining [20], decision
evaluation, and granular computing.
It is well-known that Pawlak’s rough set and most of the
expanded rough sets were constructed on the basis of one
and only one set of the information granules, which are
generated from a binary relation or a covering. From this
point of view, we may call these rough sets the single-
granulation rough sets. In single-granulation rough sets, a
partition is a granulation space, a binary neighborhood
system [6–8] induced by a binary relation is a granulation
space, a covering [27–29] is also a granulation space.
Nevertheless, it should be noticed that in Refs. [13, 15, 16],
the authors said that we often need to describe concurrently
a target concept from some independent environments, that
is, multigranulation spaces are needed in problem solving.
From this point of view, Qian et al. [14, 17, 18] proposed
the concept of the multigranulation rough sets. The main
difference between single-granulation and multigranulation
rough sets is that we can use multi-different sets of the
X. Yang (&) X. Song
School of Computer Science and Engineering,
Jiangsu University of Science and Technology,
Zhenjiang 212003, People’s Republic of China
e-mail: yangxibei@hotmail.com
X. Song
e-mail: xnsong@yahoo.com.cn
Z. Chen
College of Information Engineering,
Taiyuan University of Technology,
Taiyuan 030024, People’s Republic of China
e-mail: chenzehua@tyut.edu.cn
J. Yang
School of Computer Science and Technology,
Nanjing University of Science and Technology,
Nanjing 210094, People’s Republic of China
e-mail: yangjy@mail.njust.edu.cn
123
Int. J. Mach. Learn. & Cyber. (2012) 3:223–232
DOI 10.1007/s13042-011-0054-8