一种新的复杂网络概述算法研究
发布时间：2019-01-02 12:05
【摘要】：针对大型复杂网络相关的概述问题展开了深入系统地研究,本文重点对属性与结构的相似度进行了全面考量,由于用户具有各自的选择属性,主要是将虚拟连接与实连接进行有效的集成,一般而言,对于大型网络数据会同时把具有相同属性的节点共同放置于k个非重叠的分类上。本文主要是以属性相似度为核心,然后将节点全部置于对应的分类中,重点采用了虚拟图概念,主要是围绕属性相似度开展的,旨在较好的划分复杂网络。另外,对子分类进行调整的过程中借助了HB-图,这样可以有助于在分类结构时,对算法进行优化。该论文为了更好地加强算法的执行效率,专门提出了诸多方法对算法加以改进。也就是说,该论文中所采用的算法,能够确保用户较好地对上卷操作(Roll-up)以及下钻操作(Drill-down)加以执行,并且,围绕各粒度层面为中心,对复杂网络的概述过程展开全面的分析。实验结果表明本文提出的基于虚连接和实连接的复杂网络概述算法OCNVR算法是切实可行的,较之于其他算法而言其执行效率更加高校。
[Abstract]:This paper focuses on the comprehensive consideration of the similarity between attributes and structures, because users have their own selection attributes. The main purpose of this paper is to integrate virtual and real connections effectively. In general, for large network data, nodes with the same attributes will be placed together on k non-overlapping categories at the same time. This paper mainly takes attribute similarity as the core, then puts all nodes in the corresponding classification, and focuses on the concept of virtual graph, mainly around attribute similarity, in order to better partition the complex network. In addition, the HB- diagram is used to adjust the subclassification, which can be helpful to optimize the algorithm in the classification structure. In order to enhance the efficiency of the algorithm, this paper puts forward many methods to improve the algorithm. That is, the algorithm used in this paper can ensure that the user performs the Roll-up and Drill-down well, and is centered around the level of granularity. A comprehensive analysis of the overview process of complex networks is carried out. The experimental results show that the proposed OCNVR algorithm based on virtual connection and real connection is feasible and more efficient than other algorithms.
【作者单位】： 郑州旅游职业学院信息工程系;河南工业大学信息科学与工程学院;
【分类号】：O157.5
，
本文编号：2398465
[Abstract]:This paper focuses on the comprehensive consideration of the similarity between attributes and structures, because users have their own selection attributes. The main purpose of this paper is to integrate virtual and real connections effectively. In general, for large network data, nodes with the same attributes will be placed together on k non-overlapping categories at the same time. This paper mainly takes attribute similarity as the core, then puts all nodes in the corresponding classification, and focuses on the concept of virtual graph, mainly around attribute similarity, in order to better partition the complex network. In addition, the HB- diagram is used to adjust the subclassification, which can be helpful to optimize the algorithm in the classification structure. In order to enhance the efficiency of the algorithm, this paper puts forward many methods to improve the algorithm. That is, the algorithm used in this paper can ensure that the user performs the Roll-up and Drill-down well, and is centered around the level of granularity. A comprehensive analysis of the overview process of complex networks is carried out. The experimental results show that the proposed OCNVR algorithm based on virtual connection and real connection is feasible and more efficient than other algorithms.
【作者单位】： 郑州旅游职业学院信息工程系;河南工业大学信息科学与工程学院;
【分类号】：O157.5
，
本文编号：2398465
本文链接：https://www.wllwen.com/kejilunwen/yysx/2398465.html
