Title :
A Fast and Efficient Algorithm for Border Extraction
Author :
Lin, Zhi-Yong ; Hao, Zhi-Feng ; Yang, Xiao-Wei
Author_Institution :
South China Univ. of Technol., Guangzhou
Abstract :
For some classification methods, such as support vector machine, the border examples are useful in the classifier design. In this paper, a fast algorithm for border extraction is proposed. The algorithm is firstly derived from the linearly separated case intuitively. Then, through kernel trick, the algorithm is extended to the nonlinearly separated case. Several simulations are carried out to evaluate the algorithm´s performance, and the results show its effectiveness. Also, the effects of the parameters on the algorithm are discussed, and some suggestions on how to select the parameters are presented.
Keywords :
feature extraction; pattern classification; support vector machines; border extraction; classification methods; classifier design; kernel function; kernel trick; pattern classification; support vector machine; Cellular neural networks; Computer science; Data mining; Kernel; Nearest neighbor searches; Neural networks; Prototypes; Supervised learning; Support vector machine classification; Support vector machines; Border extraction; Kernel function; Pattern classification;
Conference_Titel :
Machine Learning and Cybernetics, 2007 International Conference on
Conference_Location :
Hong Kong
Print_ISBN :
978-1-4244-0973-0
Electronic_ISBN :
978-1-4244-0973-0
DOI :
10.1109/ICMLC.2007.4370718
