Earticle

초록

영어

Aiming at the incomplete information systems on the condition of no prior domain knowledge, several known model extension based on the rough set theory are introduced at first, such as the tolerance relation, non-symmetric similarity relation, limited tolerance relation and valued tolerance relation. Then the merits and drawbacks of several existing valued tolerance relations are compared in this article. Next, the experiments on some UCI data sets have been done ,based on the experimental result, the author discuss the relationship between the threshold selection and classification accuracy of statistical valued tolerance relation(SVT) . Directing at the difficulty of selecting a suitable threshold, the author presents a new improved valued tolerance relation (NVT) which can choose proper threshold automatically on the basis of each data set’s feature. Experiment results indicate that the new relation can get better classification accuracy than the other extension models in dealing with the incomplete system which has small incomplete degree

목차

Abstract
 1. Introduction
 2. The Comparison of Some Known Valued Tolerance Relation
  2.1. The Valued Tolerance Relation Raised By Stefanowski
  2.2. Statistical Valued Tolerance Relation (SVT)
 3. A New Improved Valued Tolerance Relation
 4. Conclusion
 Acknowledgements
 References

키워드

저자

• Liu ping [ The school of Information Engineering, Nanchang University, Nanchang 330031, China ]
• Xiong xiaoxiao [ The school of Information Engineering, Nanchang University, Nanchang 330031, China ]
• Yang Junping [ Jiangxi University of Traditional Chinese Medicine ]

참고문헌

발행기관

간행물

표지보기 관심저널 등록

• 간행물명

  International Journal of Hybrid Information Technology

• 간기

  격월간

• pISSN

  1738-9968

• 수록기간

  2008~2016

• 십진분류

  KDC 505 DDC 605