Title :
A New Multi-class Classification Algorithm of Support Vector Machine
Author :
Zhu, Yanwei ; Zhang, Yongli
Author_Institution :
Dept. of Math. & Inf. Sci., Tang Shan Teacher´´s Coll., Tang Shan, China
Abstract :
A kind of method of SVM for multi-class problems was given in this paper. This method is based on PCA Support Vector Machine coding method. After experimenting, it is better than one-against-one Method and one-against-the -rest Method. This SVM for multi-class method saves time and enhances precision of forecast. In based on principal component analysis SVM method, coding multi-class method classes´ seismic facies, the precision of forecast is very high.
Keywords :
pattern classification; principal component analysis; support vector machines; PCA; SVM; multiclass classification algorithm; one-against-one method; one-against-the-rest method; principal component analysis; support vector machine; Accuracy; Algorithm design and analysis; Classification algorithms; Encoding; Principal component analysis; Support vector machines; Training; component analysis; multi-classification; principal; seismic facies; support vector machine;
Conference_Titel :
E-Business and E-Government (ICEE), 2010 International Conference on
Conference_Location :
Guangzhou
Print_ISBN :
978-0-7695-3997-3
DOI :
10.1109/ICEE.2010.381
