|
|
|
Application of adaptive ensemble algorithm based on
correctness and diversity |
| LUO Jian-hong1,2, CHEN De-zhao1 |
1.Department of Chemical and Biological Engineering, Zhejiang University, Hangzhou 310027, China;
2.Department of Management Science and Engineering, Zhejiang SciTech University, Hangzhou 310018, China |
|
|
|
Abstract A new ensemble algorithm was proposed. to select individual classifier reasonably from ensemble classifier in order to improve the effect of ensemble learning. Designs a selective index giving consideration to correctness and diversity based on knowledge granular principle, in order to fast select some individuals from the trained classifiers to build 'classifiers space’. Then each specific ensemble classifier is generated for each class by adaptive strategy, and these ensemble classifiers are inclusive, so the group of ensemble classifiers would cost little computation resource. Then make classification decision by adaptive strategy. Experiments conducted on some typical classification problems demonstrate that compared to the other ensemble methods, this algorithm is higher efficient, more stable and has stronger generalization performance．
|
|
Published: 16 March 2012
|
|
|
兼顾正确率和差异性的自适应集成算法及应用
针对如何从集成分类器中合理地筛选个体以提高集成学习的效果这一难题,提出了新的集成算法.该算法基于知识粒原理设计一种兼顾正确率和差异性的筛选指标,以便从训练的一批分类器中快速地选择个体组建成库；以自适应方式,针对每一类别生成特定的集成分类器,这些集成分类器间存在包容性,由此构建的集成分类器组将占用较少的计算资源,并将以自适应方式进行分类决策.对多种模式分类问题的试验结果表明：与其他集成方法相比,该集成算法更为高效,稳定性更好,具有较强的泛化性能.
|
|
| ［1］ 任俊,李志能. 支持向量机在字符分类识别中的应用［J］. 浙江大学学报:工学版, 2005,39(8):1136-1141．
REN Jun, LI Zhineng. Application of support vector machines in classification and recognition of characters［J］. Journal of Zhejiang University: Engineering Science, 2005, 39(8):11361-141．
［2］ BREIMAN L. Bagging predictors ［J］. Machine Learning, 1996, 24 (2): 123-140．
［3］ MARTINEZMUNOZ G, SUAREZ A. Pruning in ordered bagging ensembles ［C］ ∥ Proceedings of the 23th International Conference on Machine Learning. Pittsburgh, PA, 2006： 609-616．
［4］ ZHOU Z H, WU J X, TANG W. Ensembling neural networks: many could be better than all ［J］. Artificial Intelligence, 2002, 137 (1/2): 239-263．
［5］ BAKKER B , HESKES T. Clustering ensembles of neural networks［J］ . Neural Networks , 2003 , 16(2) :261-269．
［6］ 傅强, 胡上序, 赵胜颖. 基于PSO算法的神经网络集成构造方法 ［J］. 浙江大学学报:工学版, 2004, 38(12):63-67．
FU Qiang, HU Shangxu, ZHAO Shengying. PSO based approach for neural network ensembles［J］. Journal of Zhejiang University: Engineering Science, 2004, 38(12): 63-67．
［7］ 苗夺谦,王国胤,刘清,等. 粒计算：过去、现在与展望 ［M］. 北京：科学出版社, 2007：143．
［8］ YAO Y Y. Granular Computing: basic issues and possible solutions ［C］ ∥ Proceedings of the Fifth Joint Conference on Information Sciences. USA: Elsevier Publishing Company, 2000: 186-189．
［9］ LAKE C L, MERZ C J. UCI repository of machine learning database［DB/OL］.［2009-10-17］. http:∥www.ics.uci.edu/～mlearn/mlrepository.html
［10］ HOUCK C R, JOINES J A, KAY, M G. A genetic algorithm for function optimization: a Matlab implementation ［R］. Raleigh： North Carolina State University, NC, 1995．
［11］ HE Y J, CHEN D Z, ZHAO W X. Ensemble classifier system based on ant colony algorithm and its application in chemical pattern classification［J］. Chemometrics and Intelligent Laboratory Systems, 2006, 80(1): 39-49．
［12］ HOPKE P K, MASSART D L. Reference data sets for chemometrical methods testing［J］. Chemometrics and Intelligent Laboratory Systems, 1993, 19(1):35-41． |
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
| |
Shared |
|
|
|
|
| |
Discussed |
|
|
|
|
