DocumentCode :
401865
Title :
The information granulation in discretization
Author :
Wang, Li-hong ; Zhang, Shu-cui ; Fan, Hui ; Wu, Geng-feng
Author_Institution :
Sch. of Comput. Sci. and Technol., Shanghai Univ., China
Volume :
5
fYear :
2003
fDate :
2-5 Nov. 2003
Firstpage :
2620
Abstract :
Discretization is a vehicle of abstraction that supports a conversion of clouds of numeric data into more tangible information granules, which can be represented as hyper-boxes in R". A partial ordering of all discretization schemes for a given information table is defined to describe the relative granularity. All discretized tables form a hierarchical structure, representing the information table under different granulation. The granule-based entropy is defined to measure the information change in discretization. With meet and join operations, the partially ordered set becomes a lattice named discretization lattice, which is a Boolean algebra.
Keywords :
Boolean algebra; data mining; learning (artificial intelligence); rough set theory; Boolean algebra; discretization lattice; discretized tables; granule based entropy; hierarchical structure; hyper boxes; information granulation; numeric data conversion; partial ordering; partially ordered set; relative granularity; Boolean algebra; Clouds; Computer science; Data analysis; Data mining; Entropy; Information analysis; Lattices; Machine learning; Vehicles;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Machine Learning and Cybernetics, 2003 International Conference on
Print_ISBN :
0-7803-8131-9
Type :
conf
DOI :
10.1109/ICMLC.2003.1259971
Filename :
1259971
Link To Document :
بازگشت