DocumentCode :
2730377
Title :
An efficient Pawlak reduction algorithm based on bitmap and granular computing
Author :
Xu, Zhangyan ; Qian, Wenbin ; Huang, Liyu
Author_Institution :
Coll. of Comput. Sci. & Inf. Eng., Guangxi Normal Univ., Guilin, China
Volume :
1
fYear :
2009
fDate :
20-22 Nov. 2009
Firstpage :
186
Lastpage :
190
Abstract :
Attribute reduction in rough set theory is an important feature selection method, which can apply successfully in data mining and machine leaning .etc. In this paper, a new algorithm of attribute reduction is proposed which based on bitmap and granular computing. At first, in order to reduce the research space, we need not to compute the record vectors if only the number of corresponding class vector is equal to one. Then a new heuristic information which can efficiently reduce the numbers of granular computing is proposed. That is to say, we reduce some computations which can not change the results of attribute reduction. At last, some different datasets on UCI are used to test the performance of the new algorithm. The experimental results show that the proposed algorithm is more efficient than the other relevant algorithms.
Keywords :
data mining; learning (artificial intelligence); rough set theory; vectors; Pawlak reduction algorithm; UCI; attribute reduction; bitmap; class vector; data mining; feature selection method; granular computing; heuristic information; machine leaning; relevant algorithms; research space; rough set theory; Algorithm design and analysis; Computer science; Data analysis; Data engineering; Data mining; Educational institutions; Heuristic algorithms; Learning systems; Set theory; Testing; attribute reduction; bitmap; granluar computing; rough set;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Intelligent Computing and Intelligent Systems, 2009. ICIS 2009. IEEE International Conference on
Conference_Location :
Shanghai
Print_ISBN :
978-1-4244-4754-1
Electronic_ISBN :
978-1-4244-4738-1
Type :
conf
DOI :
10.1109/ICICISYS.2009.5357905
Filename :
5357905
Link To Document :
بازگشت