不完备混杂信息系统的属性约简  

Attribute Reduction for an Incomplete Hybrid Information System

作　　者：王艇 李博 朱重龙 胡晶晶 WANG Ting;LI Bo;ZHU Chonglong;HU Jingjing(School of Electronic Engineering,Jiangsu Vocational College of Electronics and Information,Huaian Jiangsu 223003,China;Modern Educational Technology Center,Jiangsu Vocational College of Electronics and Information,Huaian Jiangsu 223003,China)

机构地区：[1]江苏电子信息职业学院电子工程学院,江苏淮安223003 [2]江苏电子信息职业学院现代教育技术中心,江苏淮安223003

出　　处：《电子器件》2024年第4期1148-1156,共9页Chinese Journal of Electron Devices

基　　金：江苏省高校“青蓝工程”项目;“苏电英才”项目;淮安市创新能力建设计划项目(HAP202203);2023年度江苏省高等学校基础科学(自然科学)研究面上项目(23KJD510001);2023年度江苏电子信息职业学院校级自然科学研究项目(JSEIY2023006);2024年度江苏省高等学校基础科学(自然科学)研究面上项目(24KJD520002)。

摘　　要：海量数据的复杂性、不确定性和高维性引发了属性约简的需求。对于完备决策系统,香农熵及其变体已被广泛用于评估属性的重要性。对于不完备混杂信息系统,基于信息增益的属性约简研究较少。在模糊粗糙集框架,针对不完备混杂信息系统提出了一种基于信息增益的属性约简方法,构造了适用于不同数据类型的距离函数来度量不同对象之间的距离,然后计算从每个属性中获得的信息,最后选择符合条件的属性子集。为了验证所提出方法在不完备混杂信息中的有效性,在几个公开数据集上进行了实验。实验结果表明提出的算法相较于基于增益比的属性约简算法、基于Alpha-investing流属性约简算法以及基于邻域条件互信息的交互属性约简算法在处理不完备混杂数据上是有效的。Massive data featured by complexity,uncertainty,and high dimensionality spark demand for attribute reduction.For complete decision systems,Shannon entropy and its variants have been widely used to evaluate the importance of attributes.Attribute reduction based on information gain is less studied for incomplete hybrid information systems.The framework of the fuzzy rough set is adopted and an attribute reduction method based on information gain is proposed for the incomplete hybrid information system.A distance function applicable to different data types is constructed to measure the distance between different objects.Then the information gained from each attribute is calculated.Finally,an attribute subset qualified for the new measurement is selected.The experimental results show that the proposed algorithm is effective compared with the attribute reduction algorithm based on gain ratio,Alpha-investing flow attribute reduction algorithm and interactive attribute reduction algorithm based on mutual information of neighborhood conditions in processing incomplete mixed data.

关 键 词：混杂数据 属性约简 不完备决策系统 模糊粗糙集 信息增益 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]