DocumentCode :
2224568
Title :
The research of the feature selection method based on the ECE and quantum genetic algorithm
Author :
Wei, Zhang ; Ye, Qiu
Author_Institution :
Shandong Provincial Key Lab. of Comput. Network, Shandong Comput. Sci. Center, Jinan, China
Volume :
6
fYear :
2010
fDate :
20-22 Aug. 2010
Abstract :
Feature selection method is the critical technique of the automatic text categorization. A new method of the text feature selection based on the quantum genetic algorithm is proposed in this paper. First of all, using the ECE statistical method to remove redundant features and noise features for the original feature set, Genetic algorithms are used to optimal feature subset; finally the best feature subset is obtained. In the method, the text vector is coded by quantum bit, and the chromosome is updated by the quantum rotating gate and quantum not-gate. Meanwhile, according to the characteristics of the information filtering, we consider adequately on the feature weight, text similarity and vector dimension in order to improve the fitness function. The experiment has proved that the method can reduce the dimension of text vector and improve the precision of text classification.
Keywords :
entropy; feature extraction; genetic algorithms; information filtering; quantum gates; statistical analysis; text analysis; ECE; automatic text categorization; chromosome; expected cross entropy; fitness function; information filtering; quantum genetic algorithm; quantum not-gate; statistical method; text feature selection method; Quantum mechanics; Feature Selection; Quantum Genetic Algorithm; Text Categorization;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Advanced Computer Theory and Engineering (ICACTE), 2010 3rd International Conference on
Conference_Location :
Chengdu
ISSN :
2154-7491
Print_ISBN :
978-1-4244-6539-2
Type :
conf
DOI :
10.1109/ICACTE.2010.5579390
Filename :
5579390
Link To Document :
بازگشت