Title :
The Improvment of Text Feature Selection Method Based on Key Words
Author :
Jian-Fang, Cao ; Hong-Bin, Wang
Author_Institution :
Dept. of Comput. Sci., Xinzhou Teachers Univ., Xinzhou, China
Abstract :
Vector space model is commonly used in the formal representation on text, but this approach would not highlight the features which play a key role in the text contents. An improved feature selection method based on key words was proposed, which uses text structural information and mutual information theory to extract key words on text content. Through using support vector machine (SVM) classifier to test, results showed that classification accuracy has improved significantly.
Keywords :
learning (artificial intelligence); support vector machines; text analysis; vectors; SVM classifier; formal representation; improved feature selection method; key words; mutual information theory; support vector machine classifier; text feature selection method; text structural information; vector space model; support vector machine; text classification; text feature selection; vector space model;
Conference_Titel :
Computing, Measurement, Control and Sensor Network (CMCSN), 2012 International Conference on
Conference_Location :
Taiyuan
Print_ISBN :
978-1-4673-2033-7
DOI :
10.1109/CMCSN.2012.36
