基于改进FCM聚类的复杂网络节点重要性评估方法
关键词 复杂网络 FCM算法 最短路 重要性评估
毕业设计说明书（论文）外文摘要
Title The Importance Of Assessment Methods
That Based On The Improved FCM Clustering
Complex Networks Nodes
Abstract
In this paper, though the analysis of the traditional FCM algorithm，we focus on the problem of the loss of accuracy of converting map into spatial data sets. We put forward an improved FCM clustering algorithm, and set node connectivity, node betweenness and network traffic as the indicators for assessing method. We’ll realize the importance of the complex network nodes by applying of the improved algorithm in traffic networks. Finally, the design and implementation of this method’s performance will test the simulation platform. The improved algorithm is the calculation of the distance in the FCM algorithm: the traditional FCM algorithm is based on the operation of the spatial data sets, which could not reflect the connectivity of the network node importance. In the topology with the shortest path between two points instead of the Euclidean distance to the cluster center. When accessing the importance of the node, the calculation of the maximum of the membership of each node of each cluster center. The higher the value, the more important the node is.
Keywords complex networks FCM algorithm shortest path the importance of assessment
目 次
1 引言 1
1.1 研究背景 2
1.2 研究内容及意义 2
1.3 论文内容的组织 3
2 复杂网络综述 4
2.1 小世界网络 4
2.2 规则网络 5
2.3 随机网络（ER网络） 6
2.4 无标度网络（BA网络） 6
3 复杂网络聚类分析 8
3．1 聚类分析 8
3．2 聚类分析方法 8
3．3 模糊C均值算法 9
3．4 FCM聚类算法存在的弱点 11
3．5 基于FCM聚类的算法改进 12
4 复杂网络节点重要度评估 17
4.1 介数法 17
4.2 节点删除法 19
4.3 节点融合法（节点收缩法） 20
4.4 节点孤立法 20
5 复杂交通网络节点重要度评估 21
5.1 基于FCM聚类的节点重要性评估算法 21
5.2 基于改进的FCM算法节点重要度评估 25
5.3 应用的局限性 25
6 实验仿真 26
6.1 传统FCM的应用 26
6.2 改进FCM的应用 28
6.3 介数法验证 31
结 论 33