Please wait a minute...
Front. Inform. Technol. Electron. Eng.  2016, Vol. 17 Issue (10): 973-981    DOI: 10.1631/FITEE.1601078
    
Max-margin based Bayesian classifier
Tao-cheng Hu, Jin-hui Yu
State Key Lab of CAD & CG, Zhejiang University, Hangzhou 310058, China
Download:   PDF(0KB)
Export: BibTeX | EndNote (RIS)      

Abstract  There is a tradeoff between generalization capability and computational overhead in multi-class learning. We propose a generative probabilistic multi-class classifier, considering both the generalization capability and the learning/prediction rate. We show that the classifier has a max-margin property. Thus, prediction on future unseen data can nearly achieve the same performance as in the training stage. In addition, local variables are eliminated, which greatly simplifies the optimization problem. By convex and probabilistic analysis, an efficient online learning algorithm is developed. The algorithm aggregates rather than averages dualities, which is different from the classical situations. Empirical results indicate that our method has a good generalization capability and coverage rate.

Key wordsMulti-class learning      Max-margin learning      Online algorithm     
Received: 10 March 2016      Published: 08 October 2016
CLC:  TP181  
Cite this article:

Tao-cheng Hu, Jin-hui Yu. Max-margin based Bayesian classifier. Front. Inform. Technol. Electron. Eng., 2016, 17(10): 973-981.

URL:

http://www.zjujournals.com/xueshu/fitee/10.1631/FITEE.1601078     OR     http://www.zjujournals.com/xueshu/fitee/Y2016/V17/I10/973


基于最大间隔的贝叶斯分类器

\n 概要:多分类学习中经常需要考虑在泛化性能和计算开销间进行权衡。本文提出一个生成式概率多分类器,综合考虑了泛化性和学习/预测速率。我们首先证明了我们的分类器具有最大间隔性质,这意味着对于未来数据的预测精度几乎和训练阶段一样高。此外,我们消除了目标函数中的大量的局部变元,极大地简化了优化问题。通过凸分析和概率语义分析,我们设计了高效的在线算法,与经典情形的最大不同在于这个算法使用聚集而非平均化处理梯度。实验证明了我们的算法具有很好的泛化性能和收敛速度。
\n

关键词: 多类学习,  最大间隔学习,  在线算法 
[1] Jian-ru Xue, Di Wang, Shao-yi Du, Di-xiao Cui, Yong Huang, Nan-ning Zheng. A vision-centered multi-sensor fusing approach to self-localization and obstacle perception for robotic cars[J]. Front. Inform. Technol. Electron. Eng., 2017, 18(1): 122-138.
[2] Jo?o Carneiro, Diogo Martinho, Goreti Marreiros, Paulo Novais. Intelligent negotiation model for ubiquitous group decision scenarios[J]. Front. Inform. Technol. Electron. Eng., 2016, 17(4): 296-308.
[3] Izabela Nielsen, Robert Wójcik, Grzegorz Bocewicz, Zbigniew Banaszak. Multimodal processes optimization subject to fuzzy operation time constraints: declarative modeling approach[J]. Front. Inform. Technol. Electron. Eng., 2016, 17(4): 338-347.
[4] Ya-tao Zhang, Cheng-yu Liu, Shou-shui Wei, Chang-zhi Wei, Fei-fei Liu. ECG quality assessment based on a kernel support vector machine and genetic algorithm with a feature matrix[J]. Front. Inform. Technol. Electron. Eng., 2014, 15(7): 564-573.
[5] Feng-fei Zhao, Zheng Qin, Zhuo Shao, Jun Fang, Bo-yan Ren. Greedy feature replacement for online value function approximation[J]. Front. Inform. Technol. Electron. Eng., 2014, 15(3): 223-231.
[6] Hong-xia Pang, Wen-de Dong, Zhi-hai Xu, Hua-jun Feng, Qi Li, Yue-ting Chen. Novel linear search for support vector machine parameter selection[J]. Front. Inform. Technol. Electron. Eng., 2011, 12(11): 885-896.
[7] Zhi-yong Yan, Cong-fu Xu, Yun-he Pan. Improving naive Bayes classifier by dividing its decision regions[J]. Front. Inform. Technol. Electron. Eng., 2011, 12(8): 647-657.
[8] Peng Chen, Yong-zai Lu. Extremal optimization for optimizing kernel function and its parameters in support vector regression[J]. Front. Inform. Technol. Electron. Eng., 2011, 12(4): 297-306.
[9] Zhuo-jun Jin, Hui Qian, Shen-yi Chen, Miao-liang Zhu. Convergence analysis of an incremental approach to online inverse reinforcement learning[J]. Front. Inform. Technol. Electron. Eng., 2011, 12(1): 17-24.
[10] Shen-yi Chen, Hui Qian, Jia Fan, Zhuo-jun Jin, Miao-liang Zhu. Modified reward function on abstract features in inverse reinforcement learning[J]. Front. Inform. Technol. Electron. Eng., 2010, 11(9): 718-723.