不完备模糊混合决策系统的邻域粗糙集分析方法  

Approach for incomplete fuzzy hybrid decision system on neighborhood rough set

作　　者：赵佰亭[1] 陈希军[1] 曾庆双[1] 

机构地区：[1]哈尔滨工业大学空间控制与惯性技术研究中心,哈尔滨150001

出　　处：《吉林大学学报（工学版）》2011年第3期721-727,共7页Journal of Jilin University:Engineering and Technology Edition

基　　金：国防科技预研基金项目(9140A17030207HT0150)

摘　　要：针对现实中同时具有不完备、模糊、混合属性值域决策系统的约简问题,建立了广义邻域粗糙集模型,提出了未知属性的辨别方法和基于属性重要度的约简算法。采用广义邻域关系度量不可分辨关系,通过邻域粒子逼近论域空间,是非对称相似关系、容差关系和模糊等价关系的广义化,可以直接处理同时含有名义型、数值型、模糊型、丢失型和遗漏型不完备属性的混合决策系统。依据分类一致性假设及广义邻域关系进行未知属性的辨别,讨论了噪声样本和邻域大小对分类精度的影响,给出了约简算法的具体实现。采用HitSHT数据和UCI数据库中2组数据进行了仿真试验,预测精度证明了模型的合理性及约简算法的有效性。In order to reduce the incomplete fuzzy hybrid decision systems, a generalized neighborhood rough set model is proposed. Discrimination methods of the missing values and a hybrid reduction algorithm are also developed. The model approximates an arbitrary subset in the universe with neighborhood granules, and the generalized neighborhood relations are the generalization of the asymmetric similarity relations, the tolerance relations and the fuzzy equivalence relations. The model can deal with the incomplete fuzzy hybrid decision system directly. The discrimination methods of the lost and the unrelated conditions are developed based on the assumption of the consistency classification. The influence of the noise samples and the neighborhood values on the classification accuracy is investigated. The validity and feasibility of the model and the reduction algorithm are demonstrated by experiments on HitSHT and two LICI machine learning databases.

关 键 词：人工智能 混合决策系统 邻域粗糙集 约简 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]