Large Sample Iterative Training Algorithm for Support Vector Machine
(Sanya Institute of Technology, Sanya 572022, China)
Abstract
Aiming at the large sample training problem for Support Vector Machine (SVM), a new iterative training algorithm is proposed. To build an initial training sample set, the K-means clustering algorithm is used to compress the training sample set, with each cluster centroid serving as the initial training sample set, reducing redundant information between samples to enhance training speed. To ensure the training accuracy, the resulting centroid is taken as the initial sample set, and the boundary samples and misclassified samples are added to the initial sample set with classification. And it is used as training samples for iterative training until the number of misclassified samples is stable. Here, the K-means clustering SVM iterative training algorithm can reduce the computational complexity while maintaining the training accuracy and improve the classification and training speed by optimization.
Keywords
Support Vector Machine; K-means clustering algorithm; iterative algorithm; Machine Learning
References
- [1] LI J M,ZHANG B,LIN F Z. Training Algorithms for Support Vector Machines [J].Journal of Tsinghua University,2003,43(1):120-124.
- [2] PLATT J C. Fast Training of Support Vector Machines Using Sequential Minimal Optimization [C]//Advances in Kernel Methods-Support Vector Learning.Cambridge:MIT Press,1999:185-208.
- [3] KEERTHI S S,SHEVADE S K,BHATTACHARYYA C,et al. A Fast Iterative Nearest Point Algorithm for Support Vector Machine Classifier Design [J].IEEE Transactions on Neural Networks,2000,11(1):124-136.
- [4] An Jinlong Research on Several Issues of Support Vector Machines [D]. Tianjin: Tianjin University, 2004
- [5] SCHÖLKOPF B,MIKA S,BURGES C J C,et al. Input Space versus Feature Space in Kernel-Based Methods [J].IEEE Transactions on Neural Networks,1999,10(5):1000-1017.
- [6] Wang Xiufei Research on Quantitative Identification of Adhesive Defects in Composite Materials Based on Feature Weighted Support Vector Machine [D]. Hohhot: Inner Mongolia University, 2011
- [7] Li Fei, Li Honglian Support Vector Machine Large Scale Sample Fast Training Algorithm [J]. Journal of Beijing University of Information Science and Technology: Natural Science Edition, 2012, 27 (2): 83-87
- [8] Liu Li Support Vector Machine and Its Application in Remote Sensing Image Processing [D]. Hefei: University of Science and Technology of China, 2005
- [9] Meng Yuanyuan Research and Application of Fuzzy Support Vector Machine [D]. Jinan: Shandong Normal University, 2006
- [10] Kong Rui Research on Kernel based Learning Methods and Their Applications in Facial Recognition [D]. Hefei: University of Science and Technology of China, 2004
- [11] Tian Xinmei, Wu Xiuqing, Liu Li A new SVM iterative algorithm under large sample conditions [J]. Computer Engineering, 2007 (8): 205-207
- [12] Guo Zhenkai, Song Zhaoqing, Mao Jianqin Chaos time series prediction based on improved SVMR [J]. Control Engineering, 2008 (4): 385-388
- [13] DUDA R O,HART P E,STORK D G. Pattern Classification: 2nd ed [M].New York:John Wiley & Sons,2001.
About the authors
Chen Jimao (1995-), male, Han ethnicity, from Ledong, Hainan, lecturer, undergraduate, research direction: mathematics, algorithms, and programming.