多源模糊信息系统空值估算方法研究  

Research on Null Value Estimation Method of Multi-Source Fuzzy Information System

作　　者：周燕 万里勇 ZHOU Yan;WAN Li-yong(Applied Science and Technology College of Jiangxi,Nanchang Jiangxi 330100,China;2.Jiangxi Normal University,Nanchang Jiangxi 330027,China)

机构地区：[1]江西应用科技学院,江西南昌330100 [2]江西师范大学,江西南昌330027

出　　处：《计算机仿真》2021年第4期159-163,共5页Computer Simulation

基　　金：江西省教育厅科学技术研究项目(GJJ91100)。

摘　　要：对传统多源模糊信息系统存在空值估算准确性差、数据不完备等问题,提出一种基于粗糙集理论的空值估算方法。对多源模糊系统中数据缺失、遗漏等情况进行分析,通过模糊覆盖法获得不完备信息的信任函数;利用粗糙集理论中不可分辨的等价关系、相容关系通过四元组属性计算,得到近似的拟合函数;通过属性约简处理获得关系表中与空值相关的属性值,解决系统中的空值估算问题。实验证明,经过多个数据集和数据表相比表明,所提方法得到的估算结果具有更高的准确率和有效性。Traditionally, the accuracy of null value estimation is low in multi-source fuzzy information systems. Therefore, a method to estimate the null value based on rough set theory was proposed. Firstly, the missing and omission of data in the multi-source fuzzy system was analyzed. And then, the belief function of incomplete information was obtained through the fuzzy coverage method. Based on the indiscernible equivalence relation and compatible relation in rough set theory, the approximate fitting function was obtained by calculating the attribute of four tuples. Finally, the attribute values related to null values in relational tables were obtained through the attribute reduction. Thus, the problem of null value estimation in system was solved. Experimental results prove that the proposed method has higher accuracy and higher availability.

关 键 词：不确定性 属性值 空值估算 信任函数 粗糙集 拟合函数 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]