摘要
为了处理属性值之间大多数相交但不具备包含关系的不完备区间值信息系统,定义了一种相似联系度容差关系。基于此关系建立了拓展粗糙集模型,并引入极大相容类技术提高近似精度。通过定义新的同异反可辨识矩阵,给出了基于同异反可辨识矩阵的属性约简算法。所建模型可根据不同的用户需求和数据集的分布特点对参数进行动态调整,更符合实际。数值例子验证了模型和算法的有效性和可行性。
In order to deal with the incomplete interval-valued information system in which attribute values are most intersect but not inclusion, a tolerance relation based on similarity connection degree is defined and extension of rough set model based on this relation is proposed. The accuracy of approximation is improved by maximal consistent class technology. By defining new identical-discrepancy-contrary discernibility matrix, an attribute reduction al- gorithm based on identical-discrepancy-contrary discernibility matrix is presented. According to the needs of differ- ent users and the distribution characteristics of the data set, the parameters can be dynamically adjusted and it is more realistic. Numerical examples illustrate the effectiveness and feasibility of the model and algorithm.
出处
《桂林电子科技大学学报》
2013年第2期144-148,共5页
Journal of Guilin University of Electronic Technology
基金
国家自然科学基金(61163041)
关键词
粗糙集
不完备区间值
信息系统
相似联系度
极大相容类
属性约简
rough set
incomplete interval-valued
information system
similarity connection degree
maximal consistent class
attribute reduction
