|本期目录/Table of Contents|

[1]张燕平,邹慧锦,赵姝.基于CCA的代价敏感三支决策模型[J].南京大学学报(自然科学),2015,51(2):447-452.[doi:10.13232/j.cnki.jnju.2015.02.032]
 Zhang Yanping,Zou Huijin*,Zhao Shu.Cost-sensitive three-way decisions model based on CCA [J].Journal of Nanjing University(Natural Sciences),2015,51(2):447-452.[doi:10.13232/j.cnki.jnju.2015.02.032]
点击复制

基于CCA的代价敏感三支决策模型()
     

《南京大学学报(自然科学)》[ISSN:0469-5097/CN:32-1169/N]

卷:
51
期数:
2015年第2期
页码:
447-452
栏目:
出版日期:
2015-04-01

文章信息/Info

Title:
Cost-sensitive three-way decisions model based on CCA

作者:
张燕平12 邹慧锦12赵姝12
(1. 安徽大学计算机科学与技术学院合肥230601; 2. 安徽大学计算智能与信号处理教育部重点实验室合肥230601)
Author(s):
Zhang Yanping12 Zou Huijin12*Zhao Shu12
(1. 230601 2. 230601) Cost-sensitive three-way decisions model based on CCA Zhang Yanping1,2, Zou Huijin1,2*Z Shu1,2 (1. School of Computer Science and Technology, Anhui University, Hefei, 230601, China; 2. Key Laboratory of Intelligent Computing and Signal Processing of Ministry of Education, Anhui University, Hefei, 230601, China)
关键词:
代价敏感 三支决策 构造性覆盖算法 决策粗糙集模型
Keywords:
cost-sensitive Three-way decisions Constructive Covering Algorithm decision-theoretic rough set model
分类号:
-
DOI:
10.13232/j.cnki.jnju.2015.02.032
文献标志码:
-
摘要:
随着数据挖掘和机器学习技术在实际问题中的广泛应用, 人们越来越多的发现实际分类问题通常具有代价敏感特性 . 代价敏感的分类是指在同一分类任务中错误分类的代价是不同的 . 介绍了一种基于构造性覆盖算法
的代价敏感三支决策模型, 即将代价敏感引入到基于构造性覆盖算法的三支决策模型 . 该模型根据误分类之间的大小关系来减少正、 负覆盖的个数, 从而调整三个域, 即正域、 负域和边界域的大小 . 引入代价敏感的目的是尽可能
的减少划分损失 . 实验对比了本文的模型分类结果和基于决策粗糙集的三支决策模型, 结果表明, 本文的模型分类结果稳定, 并且能够通过改变三个域的大小, 把分类损失最小化 .
Abstract:
As data mining and machine learning techniques are widely used in practical problems, we find that more and more actual classification problems typically have cost-sensitive characteristics. The cost-sensitive classification refers to the cost of classification is different in the same category classification task. This paper introduces a cost-sensitive three-way decisions model based on Constructive Covering Algorithm, i.e., connects the characteristics cost-sensitive with three-way decisions model based on Constructive Covering Algorithm. This new model reduces the number of covers and modifies the three regions according to the loss functions. The three regions are positive region, negative region and boundary region, respectively. The purpose to introduce cost-sensitive is to reduce the division cost as far as possible. Experiments compare the new model which is cost-sensitive three-way decision model based on Constructive Covering Algorithm with three-decision model based on rough set of decisions. The experimental results show that the classification performance of proposed model is stable and the new model can minimize the classification cost according to modify the size of three regions.

参考文献/References:

[1].Yao Y Y. Two semantic issues in a probabilistic rough set model. Fundamenta Informaticae, 2011, 108(3): 249~265.
[2].[3] Yao Y Y. The superiority of three-way decisions in probabilistic rough set models. Information Science, 2011, 181(6):1080~1096.
[3].[4] Yao Y Y. Two semantic issues in a probabilistic rough set model. Fundamenta Informaticae, 2011, 108(3-4): 249~265.
[4].[5] 刘盾李天瑞李华雄. 粗糙集理论:基于三支决策视角.南京大学学报(自然科学), 2013, 49():574~581.
[5].[6] Drummond C, Holte R C. Explicitly representing expected cost: An alternative to ROC representation. In: Proceeding of the 6th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Ottawa: Elsevier, 2000, 198~207.
[6].[7] 凌晓峰.代价敏感分类器的比较研究计算机学报, 2007, 30(8): 12031212.
[7].[8] Domingos P. MetaCost: A general method for making classifiers cost-sensitive. In: Proceedings of the 5th International Conference on Knowledge Discovery and Data Mining. New York, NY, USA: ACM, 1999, 155~164.
[8].[9] Elkan C. The foundations of cost-sensitive learning. In: Proceedings of 17th International Joint Conference on Artificial Intelligence.San Diego: Elsevier Inc, 2011, 973~978.
[9].[10] Chai X Y, Deng L. Test-cost sensitivenaive Bayes classification. In: Proceedings of the 4th IEEE International Conference on Data Mining. Washington, DC, USA: IEEE Computer Society, 2004, 51~58.
[10].[11] Zhou Z H, Liu X Y. Training cost-sensitive neural networks with methods addressing the class imbalance problem. IEEE Transactions on Knowledge and Data Engineering, 2006, 18(1): 63~77.
[11].[12] 龙军, 殷建平, 祝恩等. 针对入侵检测的代价敏感主动学习算法. 南京大学学报(自然科学), 2008, (5):528~534.
[12].[13] 蔫盛益,谢照青,余雯基于代价敏感的朴素贝叶斯不平衡数据分类研究计算机研究与发展, 2011, 48(Suppl.): 387390.
[13].[14] 李华雄,周献中决策粗糙集与代价敏感分类计算机科学与探索, 2013,7(2): 126135.
[14].[15] 谢骋, 商琳. 基于三支决策粗糙集的视频异常行为检测. 南京大学学报(自然科学), 2013, (4):475~482.
[15].[16] Zhou B, Yao Y Y. Cost-sensitive three-way email spam filtering. Journal of Intelligent Information Systems, 2014, 42(1): 1~27.
[16].[17] Zhang Y P, Xing H. A three-way decisions model based on constructive covering algorithm. In: Proceeding of the 8th International Conference on Rough Set and Knowledge Technology Halifax, NS, Canada: Springer, Heidelberg, 2013, 346~353.
[17].基金项目:国家自然科学基金(61175046,61203290), 安徽省高等学校省级自然科学研究项目(KJ2013A016)

相似文献/References:

备注/Memo

备注/Memo:
国家自然科学基金(61175046,61203290) ,安徽省高等学校省级自然科学研究项目(KJ2013A016),安徽大学研究学术创新强化项目( yqh100176 )
更新日期/Last Update: 2015-03-06