The Research of Extracting Minimal Decision Rules from the Decision Table in Rough Sets

Abstract:

Analyzes the traditional methods of extracting decision rules in Rough Sets, defines the concept of the decision dependability and proposes a novel algorithm of extracting short decision rules. Only the length of decision rules is extended when the current decision rules can’t classify all the samples in the decision table. At the same time, three methods are proposed to reduce the computational complexity: 1) defines the concept of bound coefficient, 2) only classify the samples with the same decision values at a time thus averting the time-consuming classification of the equivalence classes with different decision values, 3) defines the Remain set and only classify the samples in the Remain set, so the computational complexity will decrease proportional with the reduction of the samples in the Remain set. Above-mentioned methods can be used directly for incomplete information systems and have great practicability.

• Periodical: Applied Mechanics and Materials (Volumes 44-47)
• Pages: 3948-3953
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.44-47.3948
• Online since: December 2010
• Authors: Wei Pan, Yi Jia Huang, Yang Sheng Wang, Hong Ji Yang
• Keywords: Condition Attribute, Decision Dependability, Decision Rule, Rough Set (RS)

References

[1] Huajun Zeng, Yinkui Zhang translate, Machine study. Beijing: Engineering Industry Publishing House (2003).

[2] Shuopin Wen, Shengyong Qiao, et al., Data and rules extracting in incomplete decision table based on decision tree, Computer applications, 23(11) (2003) 17-22.

[3] Changsong Sun, Xiguo Dong, Jianpei Zhang, Algorithm based on rough set and decision tree to gain minimal classing rule set, Journal of Harbin Engineering University, 23(5) (2002) 87-91.

[4] Jing Yang, Hao Wang, Xue-Gang Hu, et al. A new classification algorithm based on rough set and entropy, International Conference on Machine Learning and Cybernetics, 1(2003) 364-367.

DOI: 10.1109/icmlc.2003.1264503

[5] Quinlan, J.R., C4. 5: Programs for machine learning, San Mateo, CA: Morgan Kaufmann (1993).

[6] Chen Jin, Luo Delin, Mu Fenxiang, An improved ID3 decision tree algorithm, Proceedings of 4th International Conference on Computer Science & Education, (2009) 127-130.

DOI: 10.1109/iccse.2009.5228509

[7] Witten, I. H., MacDonald, B. A., Using concept learning for knowledge acquisition, International Journal of Man-Machine Studies, 27 (1988) 349-370.

[8] Mitchell, T.M., Utgoff, P.E., Banerji, R., Learning by experimentation: acquiring and modifying problem-solving heuristics, Machine Learning, 1 (1983) 163-190.

DOI: 10.1016/b978-0-08-051054-5.50010-8

[9] Guofang Qiu, Zhaohui Zhu, Decision rules extracting and algorithms to inferences based on classic-fuzzy variable precision concept lattices, Computer Science, 36(12) (2009) 216-218.

[10] Min Li, Shaobo Deng, Qing Liu, Rule extracting method based on relative decision condition, Computer Engineering, 35(23) (2009) 175-180.

[11] Xiaoyu Song, Chunguang Chang, Feng Liu, Rough set theory based reduction algorithm for decision table, Proceedings of the Eighth International Conference on Machine Learning and Cybernetics, (2009) 2318-2323.

DOI: 10.1109/icmlc.2009.5212161

[12] Dongkai Wang, Liang Liang, Decision rules discovery of incomplete decision table, Computer Science, 28(5) (2001) 83-85.

[13] Shuopin Wen, Shengyong Qiao, et al., Data and rules extracting in incomplete decision table based on decision tree, Computer applications, 23(11) (2003) 17-22.