计算机科学 ›› 2014, Vol. 41 ›› Issue (12): 148-150.doi: 10.11896/j.issn.1002-137X.2014.12.031

• 人工智能 • 上一篇    下一篇

最小相关性最大依赖度属性约简

翟俊海,万丽艳,王熙照   

  1. 河北大学数学与计算机学院 河北省机器学习与计算智能重点实验室 保定071002;河北大学数学与计算机学院 河北省机器学习与计算智能重点实验室 保定071002;河北大学数学与计算机学院 河北省机器学习与计算智能重点实验室 保定071002
  • 出版日期:2018-11-14 发布日期:2018-11-14
  • 基金资助:
    本文受国家自然科学基金项目(71371063,0),河北省自然科学基金项目(F2013201220,F2013201110),河北省高等学校科学技术研究重点项目(ZD20131028)资助

Attribute Reduction with Principle of Minimum Correlation and Maximum Dependency

ZHAI Jun-hai,WAN Li-yan and WANG Xi-zhao   

  • Online:2018-11-14 Published:2018-11-14

摘要: 在经典粗糙集中,基于重要度的决策表属性约简算法只考虑了决策属性与条件属性之间的依赖度,没有考虑约简中条件属性之间的相关性,由此求出的约简中可能依然包含冗余属性。针对这一问题,提出了一种改进算法,它利用最小相关性和最大依赖度准则求决策表属性约简。与基于重要度的决策表属性约简算法相比,本算法求出的约简包含的属性个数少、冗余小。实验结果显示,本算法优于基于重要度的决策表属性约简算法。

关键词: 粗糙集,决策表,属性约简,最小相关性,最大依赖度

Abstract: In the classical rough set,the reduction algorithm based on significance for decision table only considers the dependency of decision attribute and condition attribute,and does not consider the correlation between the condition attributes in reduct.The reduct calculated with this kind of algorithm may include redundant attributes.In order to deal with this problem,an improved algorithm was proposed in this paper,which calculates the reduct with the principle of minimum correlation and maximum dependency.Compared with the reduction algorithm based on significance for decision table,less attributes are remained in the reducts calculated with the proposed algorithm,and the redundancy of the reduct is smaller.The experimental results show that the proposed algorithm outperforms the reduction algorithm based on significance for decision table.

Key words: Rough sets,Decision table,Attribute reduct,Minimum correlation,Maximum dependency

[1] Pawlak Z.Rough Sets [J].International Journal of Computer and Information Science,1982,11(5):341-356
[2] 苗夺谦,李道国.粗糙集理论、算法与应用 [M].北京:清华大学出版社,2008
[3] Skowron A,Rauszer C.The discernibility matrices and functions in information systems [C]∥Slowinski R,ed.Intelligent Decision Support,Handbook of Applications and Advances of the Rough Sets Theory.Kluwer,Dordrecht,1992
[4] 苗夺谦,王珏.粗糙集理论中知识粗糙性与信息熵关系的讨论[J].模式识别与人工智能,1998,11(1):34-40
[5] Qian Y,Liang J,Pedrycz W,et al.Positive approximation:an accelerator for attribute reduction in rough set theory [J].Artificial Intelligence,2010,174(9/10):597-618
[6] Chen D,Zhao S,Zhang L,et al.Sample Pair Selection for Attribute Reduction with Rough Set [J].IEEE Transactions on Knowledge and Data Engineering,2012,24(11):2080-2093
[7] Hu Q,Xie Z,Yu D.Hybrid attribute reduction based on a novel fuzzy-rough model and information granulation [J].Pattern Recognition,2007,0(12):3509-3521
[8] Tsang E,Chen D,Yeung D.Attributes reduction using fuzzyrough sets [J].IEEE Transactions on fuzzy system,2008,6(5):1130-1141
[9] Yao Y Y,Zhao Y.Discernibility matrix simplification for constructing attribute reducts [J].Information Sciences,2009,179(5):867-882
[10] Wang J,Wang J.Reduction algorithms based on discernibility matrix:the ordered attributes method [J].Journal of Computer Science and Technology,2001,16(6):489-504
[11] Qian Y,Liang J.Combination Entropy & Combination Granulation in Rough Set Theory [J].International Journal of Uncertainty,Fuzziness and Knowledge-Based Systems,2008,16(2):179-193
[12] 王钰,刘三阳,刘振华.相容 RS理论中的属性约简算法 [J].计算机科学,2003,30(2):161-162
[13] Parthalain N,Shen Q.Exploring the boundary region of tolerance rough sets for feature selection [J].Pattern Recognition,2009,42(5):655-667
[14] Jensen R,Shen Q.New Approaches to Fuzzy-Rough Feature Selection [J].IEEE Transactions on Fuzzy Systems,2009,17(4):824-838
[15] Yao Y Y,Zhao Y.Attribute reduction in decision-theoreticrough set models [J].Information Sciences,2008,178(17):3356-3373
[16] Cover T M,Thomas J A.Elements of information theory (Second Edition) [M].Hoboken:Wiley-Interscience,2006
[17] Blake C L,Merz C J.UCI Repository of Machine Learning Databases.[2011].http://www.ics.uci.edu/~mlearn/MLRepository.html
[18] Mitchell T M.Machine Learning [M].北京:机械工业出版社,2003:55-73

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!