TY - GEN
T1 - A Fast Support Vector Machine Classification Algorithm Based on Karush-Kuhn-Tucker Conditions
AU - ZHANG, Ying
AU - WANG, Xizhao
AU - ZHAI, Junhai
N1 - This research is supported by the Natural Science Foundation of Hebei Province (F2008000635), by the key project foundation of applied fundamental research of Hebei Province (08963522D), by the plan of 100 excellent innovative scientists of the first group in Education Department of Hebei Province.
PY - 2009
Y1 - 2009
N2 - Although SVM have shown potential and promising performance in classification, they have been limited by speed particularly when the training data set is large. In this paper, we propose an algorithm called the fast SVM classification algorithm based on Karush-Kuhn-Tucker (KKT) conditions. In this algorithm, we remove points that are independent of support vectors firstly in the training process, and then decompose the remaining points into blocks to accelerate the next training. From the theoretical analysis, this algorithm can remarkably reduce the computation complexity and accelerate SVM training. And experiments on both artificial and real datasets demonstrate the efficiency of this algorithm.
AB - Although SVM have shown potential and promising performance in classification, they have been limited by speed particularly when the training data set is large. In this paper, we propose an algorithm called the fast SVM classification algorithm based on Karush-Kuhn-Tucker (KKT) conditions. In this algorithm, we remove points that are independent of support vectors firstly in the training process, and then decompose the remaining points into blocks to accelerate the next training. From the theoretical analysis, this algorithm can remarkably reduce the computation complexity and accelerate SVM training. And experiments on both artificial and real datasets demonstrate the efficiency of this algorithm.
KW - Agglomerative hierarchical clustering algorithm
KW - Karush-Kuhn-Tucker (KKT) conditions
KW - Remove samples
KW - Support Vector Machine
UR - http://www.scopus.com/inward/record.url?scp=76649083624&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-10646-0_46
DO - 10.1007/978-3-642-10646-0_46
M3 - Conference paper (refereed)
AN - SCOPUS:76649083624
SN - 9783642106453
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 382
EP - 389
BT - Rough Sets, Fuzzy Sets, Data Mining and Granular Computing : 12th International Conference, RSFDGrC 2009, Proceedings
A2 - SAKAI, Hiroshi
A2 - CHAKRABORTY, Mihir Kumar
A2 - HASSANIEN, Aboul Ella
A2 - ŚLĘZAK, Dominik
A2 - ZHU, William
PB - Springer-Verlag, Berlin, Heidelberg
T2 - 12th International Conference on Rough Sets, Fuzzy Sets, Data Mining and Granular Computing, RSFDGrC 2009
Y2 - 15 December 2009 through 18 December 2009
ER -