For Full-Text PDF, please login, if you are a member of IEICE,|
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
On the Properties and Applications of Inconsistent Neighborhood in Neighborhood Rough Set Models
Shujiao LIAO Qingxin ZHU Rui LIANG
IEICE TRANSACTIONS on Information and Systems
Publication Date: 2018/03/01
Online ISSN: 1745-1361
Type of Manuscript: PAPER
Category: Artificial Intelligence, Data Mining
inconsistent neighborhood, neighborhood rough set, properties, attribute reduction, fast forward algorithm, run-time,
Full Text: PDF(753.3KB)
>>Buy this Article
Rough set theory is an important branch of data mining and granular computing, among which neighborhood rough set is presented to deal with numerical data and hybrid data. In this paper, we propose a new concept called inconsistent neighborhood, which extracts inconsistent objects from a traditional neighborhood. Firstly, a series of interesting properties are obtained for inconsistent neighborhoods. Specially, some properties generate new solutions to compute the quantities in neighborhood rough set. Then, a fast forward attribute reduction algorithm is proposed by applying the obtained properties. Experiments undertaken on twelve UCI datasets show that the proposed algorithm can get the same attribute reduction results as the existing algorithms in neighborhood rough set domain, and it runs much faster than the existing ones. This validates that employing inconsistent neighborhoods is advantageous in the applications of neighborhood rough set. The study would provide a new insight into neighborhood rough set theory.