摘要
为了解决核事故应急决策中属性冗余并可能干扰正确评估方案的问题,提出一种基于广义变精度模糊粗糙集的属性约简算法。首先将每个方案对应的属性值量化,根据量化后的值建立模糊相似类,引入误差参数和相对错误分类率求出不同属性组合下的近似分类质量进行属性约简,并计算约简集合下属性的客观权重,将被评方案在每个约简集合下到理想点的贴近度加权求和以确定其优选顺序。实例分析证明了该算法的有效性。
To solve the problem of redundant attributes that may interfere with program evaluation in nuclear accident emergency decision,an attribute reduction algorithm based generalized variable precision fuzzy-roughness sets is proposed.Firstly,the algorithm transforms the real attribute value into quantitative value,establishes fuzzy similarity classes according to the latter value,introduces error parameter and relative misclassification rate to calculate approximate classification quality to achieve reduced attributes and their objective weights by the attribute importance,and then constructs the approximation degree between the programs and ideal points under each set of reduced attributes to select the most optimal one.At last,an example proves the validity of this method.
出处
《辐射防护》
CAS
CSCD
北大核心
2011年第2期100-104,共5页
Radiation Protection
基金
国家自然科学基金项目(11065002)
江西省自然科学基金项目(2010GZW0001)
关键词
属性约简
核事故应急决策
粗糙集
模糊相似类
优选
Attribute Reduction
Nuclear Accident Emergency Decision
Roughness Set
Fuzzy Similarity Classes
Optimization
