一种用于方向预测的集成学习算法

摘要/Abstract

摘要： 对用方向向量标识示例的学习问题，以预测方向与实际方向之间的方向误差最小化为目标，提出了一种可用于方向预测的集成学习算法，详细分析了构造多个预测函数以及组合各个预测函数以实现方向的最优化预测方法. 提出的算法具有广泛的应用特性：当用不同的轴向来标识类别时，可简化得到多分类连续AdaBoost算法，其能确保训练错误率随分类器个数增加而降低；用错分代价组成的向量来标识示例时，可简化得到一种平均错分代价最小化的集成学习算法. 理论分析和实验结果均表明了算法的合理性和有效性.

关键词: , 结构化预测, 方向预测, 模糊分类, 代价敏感, AdaBoost算法

Abstract: To resolve the learning problem in which the instances are labeled by vectors, with the destination of direction error minimization between the direction represented by prediction vector and the direction represented by actual vector, an ensemble learning algorithm for direction prediction was proposed. The methods to construct multiple prediction functions and to combine them to realize the optimized prediction of instance directions were put forward. This algorithm is very general. When the different classes are labeled by the different direction vectors of axes, the proposed algorithm is degenerated to real AdaBoost algorithm for multiclass classification, guaranteeing that the training error of the combination classifier can
be reduced while the number of trained classifiers increases. When the instances are labeled by the vector composed of the classification costs of all classes, the proposed algorithm is degenerated to an ensemble learning algorithm for costsensitive classification which can minimize average classification cost. The theoretical analysis and experimental results show that the proposed algorithm is reasonable and effective.

Key words: structured prediction, direction prediction, fuzzy classification, costsensitive, AdaBoost algorithm

中图分类号:

TP 391

付忠良. 一种用于方向预测的集成学习算法 [J]. 上海交通大学学报（自然版）, 2012, 46(02): 250-058.

FU Zhong-Liang. An Ensemble Learning Algorithm for Direction Prediction[J]. Journal of Shanghai Jiaotong University, 2012, 46(02): 250-058.

参考文献

［1］Lafferty J, McCallum A, Pereira F. Conditional random fields: Probabilistic models for segmenting and labeling sequence data ［C］//Proc of the International Conf on Machine Learning. San Mateo: Morgan Kaufmann, 2001:282289.
［2］Taskar B, Guestrin C, Koller D. Maxmargin Markov networks ［C］//Advances in Neural Information Processing Systems16. Cambridge: MIT, 2004:2532.
［3］Tsochantaridis I, Hofmann T, Joachims T, et al. Support vector machine learning for interdependent and structured output spaces ［C］//Proc of the International Conf on Machine Learning. Menlo Park: AAAI, 2004:18.
［4］叶少珍,张钹,吴鸣锐,等. 一种基于神经网络覆盖构造法的模糊分类器［J］. 软件学报, 2003, 14(3): 429434.
YE Shaozhen, ZHANG Bo, WU Mingrui, et al. A fuzzy classifier based on the constructive covering approach in neural networks ［J］. Journal of Software, 2003, 14(3): 429 434.
［5］阳爱民.模糊分类模型的研究［D］. 上海:复旦大学计算机与信息技术系, 2005.
［6］Freund Y, Schapire R E. A decisiontheoretic generalization of online learning and an application to boosting ［J］. Journal of Computer and System Sciences, 1997, 55(1): 119139.
［7］Schapire R E, Singer Y. Improved boosting algorithms using confidencerated predictions ［J］. Machine Learning, 1999, 37(3): 297336.
［8］Zhu J, Rosset S, Zou H, et al. Multiclass AdaBoost ［J］. Statistics and Its Interface, 2009(2): 349360.
［9］付忠良. 分类器线性组合的有效性和最佳组合问题的研究［J］. 计算机研究与发展, 2009, 46(7): 12061216.
FU Zhongliang. Effictive property and best combination of classifiers linear combination ［J］. Journal of Computer Research and Development, 2009, 46(7): 12061216.
［10］Boutell M R, Luo J, Shen X P, et al. Learning multilabel scene classification ［J］. Pattern Recognition, 2004, 37: 17571771.
［11］Zhou Z H, Zhang M L. Multiinstance multilabel learning with application to scene classification ［C］ //Advances in Neural Information Processing Systems 19. Cambridge: MIT, 2007:16091616.
［12］Ling C X, Sheng V S. A comparative study of costsensitive classifiers ［J］. Chinese Journal of Computers, 2007, 30(8): 12031211.
［13］Ting K M, Zheng Z. Boosting costsensitive trees ［C］ //Proc of the 1st International Conf on Discovery Science. London: Springer, 1998: 244255.
［14］付忠良. 关于AdaBoost有效性的分析［J］. 计算机研究与发展, 2008, 45(10): 17471755.
FU Zhongliang. The effectiveness analysis of AdaBoost ［J］. Journal of Computer Research and Development, 2008, 45(10): 17471755.
［15］付忠良. 多分类问题代价敏感AdaBoost算法［J］. 自动化学报, 2011, 37(8): 973983.
FU Zhongliang. Costsensitive AdaBoost algorithm for multiclass classification problems ［J］. Acta Automatica Sinica, 2011, 37(8): 973983.
［16］Fan W, Stolfo S J, Zhang J, et al. AdaCost: Misclassification costsensitive boosting ［C］ //Proc of the 16th International Conf on Machine Learning. San Mateo: Morgan Kaufmann, 1999: 97105.
［17］Blake C, Keogh E, Merz C. UCI repository of machine learning databases ［EB/OL］. （19980603）［20110314］. http://www.ics.uci.edu/~mlearn/MLRepository.html.

[1]	王聚团, 戚晓宁, 黄志明. 水下生产管汇测试技术及其改进研究[J]. 海洋工程装备与技术, 2022, 9(2): 43-49.
[2]	袁振钦, 邹科, 孙亚峰, 刘刚, 屈衍, 李居跃. 基于时域分析法的动态电缆疲劳分析[J]. 海洋工程装备与技术, 2022, 9(2): 50-55.
[3]	王娟, 杨明旺, 郑茂尧, 刘凌云, 赵立君. 高强钢在大型半潜式平台组块建造中的应用[J]. 海洋工程装备与技术, 2022, 9(1): 27-31.
[4]	陈欣, 赵晓磊, 王立坤, 肖德明, 张腾月. 深水大型吸力锚建造技术研究[J]. 海洋工程装备与技术, 2022, 9(1): 32-36.
[5]	尹彦坤, 易涤非. 半潜式生产平台船体结构关键节点工程临界评估[J]. 海洋工程装备与技术, 2022, 9(1): 52-57.
[6]	ZHANG Shengfa (张胜发), TANG Na (唐纳), SHEN Guofeng (沈国峰), WANG Han (王悍), QIAO Shan (乔杉). Universal Software Architecture of Magnetic Resonance-Guided Focused Ultrasound Surgery System and Experimental Study[J]. J Shanghai Jiaotong Univ Sci, 2021, 26(4): 471-481.
[7]	MA Qunsheng (马群圣), CEN Xingxing (岑星星), YUAN Junyi (袁骏毅), HOU Xumin (侯旭敏). Word Embedding Bootstrapped Deep Active Learning Method to Information Extraction on Chinese Electronic Medical Record[J]. J Shanghai Jiaotong Univ Sci, 2021, 26(4): 494-502.
[8]	KONG Xiangqiang (孔祥强), MENG Xiangxi (孟祥熙), LI Jianbo (李见波), SHANG Yanping (尚燕平), CUI Fulin (崔福林) . Comparative Study on Two-Stage Absorption Refrigeration Systems with Different Working Pairs[J]. J Shanghai Jiaotong Univ Sci, 2021, 26(2): 155-162.
[9]	ZHUANG Weimin (庄蔚敏), WANG Pengyue (王鹏跃), AO Wenhong (熬文宏), CHEN Gang (陈刚) . Experiment and Simulation of Impact Response of Woven CFRP Laminates with Different Stacking Angles[J]. J Shanghai Jiaotong Univ Sci, 2021, 26(2): 218-230.
[10]	安庆升, 孙立东, 武秋生. 碳纤维增强复合材料发射筒设计研究[J]. 空天防御, 2021, 4(2): 13-.
[11]	ZHOU Xuhui (周旭辉), ZHANG Wenguang (张文光), XIE Jie (谢颉). Effects of Micro-Milling and Laser Engraving on Processing Quality and Implantation Mechanics of PEG-Dexamethasone Coated Neural Probe[J]. J Shanghai Jiaotong Univ Sci, 2021, 26(1): 1-9.
[12]	HUANG Ningning (黄宁宁), MA Yixin (马艺馨), ZHANG Mingzhu (张明珠), GE Hao (葛浩), WU Huawei (吴华伟). Finite Element Modeling of Human Thorax Based on MRI Images for EIT Image Reconstruction[J]. J Shanghai Jiaotong Univ Sci, 2021, 26(1): 33-39.
[13]	WANG Xianjin, GAO Xu, YU Kuigang . Fixture Locating Modelling and Optimization Research of Aluminum Alloy Sidewall in a High-Speed Train Body[J]. J Shanghai Jiaotong Univ Sci, 2020, 25(6): 706-713.
[14]	QIAO Xing, MA Dan, YAO Xuliang, FENG Baolin. Stability and Numerical Analysis of a Standby System[J]. J Shanghai Jiaotong Univ Sci, 2020, 25(6): 769-778.
[15]	WU Jin, MIN Yu, YANG Xiaodie, MA Simin . Micro-Expression Recognition Algorithm Based on Information Entropy Feature[J]. Journal of Shanghai Jiao Tong University(Science), 2020, 25(5): 589-599.