[1] Sun Liang, Han Chongzhao,. Knowledge discovery methodfor feature-decision level fusion of multiple classifiers [J]. Journal of Southeast University (English Edition), 2006, 22 (2): 222-227. [doi:10.3969/j.issn.1003-7985.2006.02.017]
Copy
Knowledge discovery methodfor feature-decision level fusion of multiple classifiers(
)
特征-决策层多分类器融合的知识发现方法
)Journal of Southeast University (English Edition)[ISSN:1003-7985/CN:32-1325/N]
- Volumn:
- 22
- Issue:
- 2006 2
- Page:
- 222-227
- Research Field:
- Computer Science and Engineering
- Publishing date:
- 2006-06-30
Info
- Title:
- Knowledge discovery methodfor feature-decision level fusion of multiple classifiers
- 特征-决策层多分类器融合的知识发现方法
- Author(s):
- Sun Liang1, 2, Han Chongzhao1
-
1 School of Electronic and Information Engineering, Xi’an Jiaotong University, Xi’an 710049, China
2 Department of Electrical Information Engineering, Institute of Information Science and Technology, Zhengzhou, 450001, China
- Keywords:
- multiple classifier fusion; knowledge discovery; Dempster-Shafer theory; generalized rough set; hyperspectral
- PACS:
- TP391
- DOI:
- 10.3969/j.issn.1003-7985.2006.02.017
- Abstract:
- To improve the performance of the multiple classifier system, a new method of feature-decision level fusion is proposed based on knowledge discovery.In the new method, the base classifiers operate on different feature spaces and their types depend on different measures of between-class separability.The uncertainty measures corresponding to each output of each base classifier are induced from the established decision tables(DTs)in the form of mass function in the Dempster-Shafer theory(DST).Furthermore, an effective fusion framework is built at the feature-decision level on the basis of a generalized rough set model and the DST.The experiment for the classification of hyperspectral remote sensing images shows that the performance of the classification can be improved by the proposed method compared with that of plurality voting(PV).
- 为进一步提高多分类器系统的分类性能, 提出了一种基于知识发现的特征-决策层多分类器融合新方法.各分类器工作于具有互补分类信息的不同特征空间且其类型由不同的类间可分性度量决定.各分类器输出的不确定性度量从建立的多个决策表中导出, 并具有条件mass函数的形式.进而基于广义粗集模型和Dempster-Shafer理论(DST)构造了一种新颖的特征-决策层融合框架.高光谱遥感图像的分类实验表明, 与多数表决融合(PV)相比, 所提出的方法可有效提高多分类器系统的分类性能.
References:
[1] Kittler J, Hatef M, Duin RPM, et al.On combining classifiers [J].IEEE Trans Pattern Analysis and Machine Intelligence, 1998, 20(3):226-239.
[2] Parhami B.Voting:a paradigm for adjudication and data fusion in dependable systems [A].In:Diab H B, Zomaya A Y, eds.Dependable Computing System [C].Wiley & Sons, Inc, 2005.
[3] Skowron A, Graymala-Busse J.From rough set to evidence theory [A]. In: Yager R R, Kacprzyk J, Fedrizzi M, eds.Advances in the Dempster-Shafer Theory of Evidence [C].New York:Wiley, 1994. 193-236.
[4] Pawlak Z.Rough set:theoretical aspects of reasoning about data [M].Boston:Kluwer Academic Publishers, 1991.
[5] Yao Y Y.Generalized rough set models [A].In:Polkowski L, Skowron A, eds.Rough Sets in Knowledge Discovery [C].Heidelberg:Physica-Verlag, 1998.286-318.
[6] Slowinski R, Vanderpooten D.A generalized definition of rough approximations based on similarity [J].IEEE Trans Knowledge and Data Engineering, 2000, 12(2):331-336.
[7] Wu W Z, Leung Y, Zhang W X.Connections between rough set theory and Dempster-Shafer theory of evidence [J].International Journal of General System, 2002, 31(4):405-430.
[8] Richards J A, Jia X P.Remote sensing digital image analysis:an introduction[M].3rd ed.Berlin, Springer-Verlag, 1999.322;331.
[9] Ma Y, Chandler J S, Wilkins D C.On the decision making problem in Dempster-Shafer theory[A].In:Proc of the International Symposium on Artificial Intelligence [C].Cancun, Mexico, 1992.7-11.
[10] William D P.The derivative ratio algorithm:avoiding atmospheric effects in remote sensing [J].IEEE Trans Geoscience and Remote Sensing, 1991, 29(3):350-357.
Memo
- Memo:
- Biographies: Sun Liang(1961—), male, graduate, associate professor, sun-liang@people.com.cn;Han Chongzhao(1943—), male, professor, czhan@mail.xjtu.edu.cn.