正在加载图片...
第15卷第6期 智能系统学报 Vol.15 No.6 2020年11月 CAAI Transactions on Intelligent Systems Nov.2020 D0L:10.11992tis.202006043 基于信息熵的对象加权概念格 张晓鹤,陈德刚',米据生2 (1.华北电力大学控制与计算机工程学院,北京102200,2.河北师范大学数学科学学院,河北石家庄 050024) 摘要:在大数据时代,由于数据规模越来越大,导致构造概念格的难度越来越高。在能够客观反映数据隐藏 信息的前提下需删除冗余对象及属性,降低数据规模,构造更为简单的概念格,从而便于用户更高效地获取知 识。为避免主观因素,本文由形式背景中属性的信息嫡来获取单属性权重,采用均值方法计算对象权重,并用 标准差计算对象重要性偏差值。通过设定的属性权重、对象权重和对象重要度偏差阈值,构造对象加权概念 格。通过实例验证了,该方法可有效删除冗余概念,简化概念格构造过程。 关键词:形式背景;概念;信息嫡:粒计算:概念格;决策规则:权值:数据挖掘 中图分类号:TP18:O236文献标志码:A 文章编号:1673-4785(2020)06-1097-07 中文引用格式:张晓鹤,陈德刚,米据生.基于信息熵的对象加权概念格JL智能系统学报,2020,15(6):1097-1103. 英文引用格式:ZHANG Xiaohe,,CHEN Degang,.MI Jusheng.Object--weighted concept lattice based on information entropyJ. CAAI transactions on intelligent systems,2020,15(6):1097-1103. Object-weighted concept lattice based on information entropy ZHANG Xiaohe',CHEN Degang',MI Jusheng (1.School of Control and Computer Engineering,North China Electric Power University,Beijing 102200,China;2.College of Math- ematics and Information Science.Hebei Normal University.Shijiazhuang 050024.China) Abstract:In the era of big data,it is becoming increasingly difficult to construct concept lattices due to the increasingly large scale of data.To objectively reflect hidden information,redundant objects and attributes should be deleted and data size should be reduced to construct simple concept lattices,thus,facilitating users to acquire knowledge efficiently.In this study,to prevent subjective factors,the information entropy of an attribute in the formal context is used to obtain a single attribute weight and the attribute weight of the object is,then,calculated using the mean value method and the im- portance deviation of the object is calculated by standard deviation.By setting the attribute weight,object weight,and object importance deviation threshold,an object-weighted concept lattice is constructed.An example is provided to veri- fy the effectiveness of this method in removing redundant concepts and simplifying the construction of concept lattices. Keywords:formal context;context,information entropy;granular computing;concept lattice;decision rules,weight value;data mining 概念格理论四是在形式背景中进行数据分析 了保持概念格的粒结构不变的属性约简。米据生 的一个重要工具,由Wille教授在1982年提出,从 等分别从变精度2]和公理化1角度考虑概念格 本质上描述了对象集和属性集的内在联系。 约简问题。针对大型数据集,陈锦坤等41将图 目前,概念格理论已在信息检索2)、数据挖 论理论与概念格理论进行结合,提出了一种快速 掘s刀、机器学习s0等领域取得了广泛应用。吴 属性约简方法。邵明文等从形式概念分析理 伟志等将粒计算与概念格理论进行结合,研究 论角度提出一种从信息表中提取决策规则的方 法。李金海等针对决策形式背景提出了一种 收稿日期:2020-06-24. 基金项目:国家自然科学基金项目(12071131,62076088). 新的知识认知和约简框架,并给出约简算法,进 通信作者:陈德刚.E-mail:zxhzxh93@126.com 一步提出了保持由全体对象集构造的决策规则不DOI: 10.11992/tis.202006043 基于信息熵的对象加权概念格 张晓鹤1 ,陈德刚1 ,米据生2 (1. 华北电力大学 控制与计算机工程学院,北京 102200; 2. 河北师范大学 数学科学学院,河北 石家庄 050024) 摘 要:在大数据时代,由于数据规模越来越大,导致构造概念格的难度越来越高。在能够客观反映数据隐藏 信息的前提下需删除冗余对象及属性,降低数据规模,构造更为简单的概念格,从而便于用户更高效地获取知 识。为避免主观因素,本文由形式背景中属性的信息熵来获取单属性权重,采用均值方法计算对象权重,并用 标准差计算对象重要性偏差值。通过设定的属性权重、对象权重和对象重要度偏差阈值,构造对象加权概念 格。通过实例验证了,该方法可有效删除冗余概念,简化概念格构造过程。 关键词:形式背景;概念;信息熵;粒计算;概念格;决策规则;权值;数据挖掘 中图分类号:TP18;O236 文献标志码:A 文章编号:1673−4785(2020)06−1097−07 中文引用格式:张晓鹤, 陈德刚, 米据生. 基于信息熵的对象加权概念格 [J]. 智能系统学报, 2020, 15(6): 1097–1103. 英文引用格式:ZHANG Xiaohe, CHEN Degang, MI Jusheng. Object-weighted concept lattice based on information entropy[J]. CAAI transactions on intelligent systems, 2020, 15(6): 1097–1103. Object-weighted concept lattice based on information entropy ZHANG Xiaohe1 ,CHEN Degang1 ,MI Jusheng2 (1. School of Control and Computer Engineering, North China Electric Power University, Beijing 102200, China; 2. College of Math￾ematics and Information Science, Hebei Normal University, Shijiazhuang 050024, China) Abstract: In the era of big data, it is becoming increasingly difficult to construct concept lattices due to the increasingly large scale of data. To objectively reflect hidden information, redundant objects and attributes should be deleted and data size should be reduced to construct simple concept lattices, thus, facilitating users to acquire knowledge efficiently. In this study, to prevent subjective factors, the information entropy of an attribute in the formal context is used to obtain a single attribute weight and the attribute weight of the object is, then, calculated using the mean value method and the im￾portance deviation of the object is calculated by standard deviation. By setting the attribute weight, object weight, and object importance deviation threshold, an object-weighted concept lattice is constructed. An example is provided to veri￾fy the effectiveness of this method in removing redundant concepts and simplifying the construction of concept lattices. Keywords: formal context; context; information entropy; granular computing; concept lattice; decision rules; weight value; data mining 概念格理论[1] 是在形式背景中进行数据分析 的一个重要工具,由 Wille 教授在 1982 年提出,从 本质上描述了对象集和属性集的内在联系。 目前,概念格理论已在信息检索[2-4] 、数据挖 掘 [5-7] 、机器学习[8-10] 等领域取得了广泛应用。吴 伟志等[11] 将粒计算与概念格理论进行结合,研究 了保持概念格的粒结构不变的属性约简。米据生 等分别从变精度[12] 和公理化[13] 角度考虑概念格 约简问题。针对大型数据集,陈锦坤等[14-15] 将图 论理论与概念格理论进行结合,提出了一种快速 属性约简方法。邵明文等[16] 从形式概念分析理 论角度提出一种从信息表中提取决策规则的方 法。李金海等[17-18] 针对决策形式背景提出了一种 新的知识认知和约简框架,并给出约简算法,进 一步提出了保持由全体对象集构造的决策规则不 收稿日期:2020−06−24. 基金项目:国家自然科学基金项目 (12071131,62076088). 通信作者:陈德刚. E-mail:zxhzxh93@126.com. 第 15 卷第 6 期 智 能 系 统 学 报 Vol.15 No.6 2020 年 11 月 CAAI Transactions on Intelligent Systems Nov. 2020
向下翻页>>
©2008-现在 cucdc.com 高等教育资讯网 版权所有