基于粗糙集的突发事件属性约简方法

doi:10.12005/orms.2018.0014

摘要/Abstract

摘要： 针对突发事件不完备信息系统中的原始数据存在大量属性冗余的问题,提出一种基于粗糙集的不完备信息系统属性约简方法,以剔除冗余属性,提高知识清晰度。首先对缺失、冗余、噪声以及连续型数据进行预处理;然后进行属性分类,将属性分为条件属性与决策属性,进而建立决策表;最后根据决策表的特征,结合有序加权平均算子的思想,提出一种基于属性重要度的启发式属性约简算法。文末,通过实例验证了方法的正确性与有效性,并利用该方法实现了火灾数据的属性约简。

关键词: 突发事件, 属性约简, 粗糙集, 不完备信息系统, 属性重要度

Abstract: With the improvement of our emphasis on emergencies and degree of informationization, large amounts of raw data in emergencies are accumulated, which contain much valuable knowledge and rules to be excavated. As emergencies involve a lot of objects in affected area, the number of attributes in raw data related to the emergency is considerable. Since there are many redundant attributes, the relationships among the factors are obscure. In addition, the missing values we cannot get lead to the incompleteness of emergency information system. For the problem of serious redundancy of attributes in raw data of emergency incomplete information systems, this paper proposes a method of attribute reduction for incomplete information systems of emergency based on rough set to eliminate redundant attributes and improve the definition of knowledge. Firstly, raw data are prepossessed. Missing data are reserved by using “*” instead. Continuous data are discretized considering the specific feature of the attribute in emergency. Redundant data and noisy data are identified by building a similarity matrix of all objects in the information system and then deleted. Then, the attributes are classified to two categories, one is conditional attribute, the other is decision attribute. On this base, we establish a decision table. Next, a method to measure the significance of attributes, which reflects the contribution of an attribute to identify the objects, is proposed combining with the ideas of the ordered weighted averaging operator. By using significance of attributes as heuristic information, we propose a heuristic attribute reduction algorithm. At the end of this paper, two examples are conducted. The first one is a classical decision table. The result shows that the algorithm in this research is correct and effective. The second one is an information table of fire events which is processed by the method in this paper. Through the second example, we show the application of the method and analyze the reduction of the information table.

Key words: emergency, attribute reduction, rough set, incomplete information system, attribute significance

中图分类号:

TP18

仲秋雁, 王然, 曲毅. 基于粗糙集的突发事件属性约简方法[J]. 运筹与管理, 2018, 27(1): 89-95.

ZHONG Qiu-yan, WANG Ran, QU Yi. A Method of Attribute Reduction for Emergency Based on Rough Set[J]. Operations Research and Management Science, 2018, 27(1): 89-95.

参考文献

[1] 常犁云,王国胤,吴渝.一种基于Rough Set理论的属性约简及规则提取方法[J].软件学报,1999,10(11):1206-1211.
[2] 马希骜,王国胤,于洪.决策域分布保持的启发式属性约简方法[J].软件学报,2014,25(8):1761-1780.
[3] Zhang X, Mei C, Chen D, et al. Multi-confidence rule acquisition and confidence-preserved attribute reduction in interval-valued decision systems[J]. International Journal of Approximate Reasoning, 2014, 55(8): 1787-1804.
[4] Wang C, Shao M, Sun B, et al. An improved attribute reduction scheme with covering based rough sets[J]. Applied Soft Computing, 2015, 26: 235-243.
[5] 梁继林,张强.不完备模糊目标信息系统的α,β精度约简算法[J].运筹与管理,2011,20(1):1-7.
[6] Qian Y, Liang J, Pedrycz W, et al. An efficient accelerator for attribute reduction from incomplete data in rough set framework[J]. Pattern Recognition, 2011, 44(8): 1658-1670.
[7] Sun L, Xu J, Tian Y. Feature selection using rough entropy-based uncertainty measures in incomplete decision systems[J]. Knowledge-Based Systems, 2012, 36: 206-216.
[8] Shu W, Shen H. Incremental feature selection based on rough set in dynamic incomplete data[J]. Pattern Recognition, 2014, 47(12): 3890-3906.
[9] 高田,杜军平,王肃.基于粗糙集的旅游突发事件属性约简[J].东南大学学报:自然科学版,2009,39(9):163-167.
[10] 吴雪莲,孙丙宇,李文波,等.基于粗糙集和CBR的救灾口粮需求预测[J].计算机工程,2012,38(9):158-161.
[11] 于峰,李向阳,李相华,等.热带气旋影响下的电网应急案例框架构建研究[J].中国安全生产科学技术,2013,9(10):78-84.
[12] 仲秋雁,郭艳敏,王宁.基于知识元的情景生成中承灾体实体化约束模型[J].系统工程,2012,30(5):75-80.
[13] Pawlak Z. Rough sets[J]. International Journal of Computer & Information Sciences, 1982, 11(5): 341-356.
[14] 王国胤.Rough集理论在不完备信息系统中的扩充[J].计算机研究与发展,2002,39(10):1238-1243.
[15] 黄海,王国胤,吴渝.一种不完备信息系统的直接约简方法[J].小型微型计算机系统,2005,26(10):1761-1765.
[16] 杨霁琳,周玉华,秦克云.不完备信息系统中属性的重要度及约简方法[J].计算机工程与应用,2010, 46(1):99-102.
[17] Yager R R. On ordered weighted averaging aggregation operators in multicriteria decisionmaking[J]. Systems, Man and Cybernetics, IEEE Transactions on, 1988, 18(1): 183-190.
[18] Kryszkiewicz M. Rough set approach to incomplete information systems[J]. Information Sciences, 1998, 112(1): 39-49.
[19] 王炜,徐章艳,王帅,等.不完备信息系统属性约简的矩阵算法[J].计算机工程,2011,37(14):36-38.
[20] 颜艳,丁健,管雪珍.不完备信息系统的一种属性约简[J].计算机工程与应用,2010,46(36):165-167.
[21] 李晓瑜,徐章艳,谭宗凤,等.基于条件分布信息量的不完备决策表属性约简[J].计算机工程与应用,2012, 48(11):107-110.
[22] 李杰,毛晓健.空气湿度与火灾发生率[J].山东消防,2003, 2:31.
[23] 葛美玲,封志明.中国人口分布的密度分级与重心曲线特征分析[J].地理学报,2009,64(2):202-210.