论文部分内容阅读
基于非负张量分解方法(NTF方法),依托2004年1月至2012年12月江苏省高速公路网络的三维OD海量交通流数据,展开网络时空特征的解析、提取与挖掘,尝试解决现有数据分析方法无法有效解析交通流网络时空演化的动态性与多维性问题。结果表明:1重构网络对原始网络的空间构型与格局具有很好的重现能力,基本刻画了原始网络的倒“不”字型空间结构;2分解的网络可提取出基本不变型、渐变型和突变型3类时变规律,且每类时间特征有各自耦合对应的局部空间格局,体现内在组织的时空统一性;3倒“不”字型空间结构由分解的沿沪宁线横向子网络叠加过润扬大桥沿扬溧线、过江阴大桥沿京沪线和过苏通大桥沿沈海线的三大纵向子网络等具有明确地理含义的多个局部空间系统共同构成,体现全局由局部组成的特性。
Based on the non-negative tensor decomposition method (NTF method), relying on the 3D OD mass flow data from January 2004 to December 2012 in Jiangsu Expressway Network, the paper analyzes and extracts the network spatio-temporal features and tries to solve the existing problems The data analysis method can not effectively analyze the dynamic and multi-dimensional issues of the temporal and spatial evolution of traffic flow networks. The results show that: 1 Reconstruction Network has a very good ability to reproduce the original network’s spatial configuration and pattern, and basically depicts the inverted “not ” font space structure of the original network; the 2 decomposition network can extract basic The three types of time-varying regularity of variation, gradual change and abrupt change, and the temporal characteristics of each type have the corresponding local spatial pattern, which reflects the spatial and temporal unification of inner organization. The horizontal sub-networks of Hu-Ning Line superimposed a number of local space systems with well-defined geographical meanings such as the three vertical sub-networks along the Yang-Li River along the Yang-Jin Bridge and the Beijing-Shanghai Line across the Yin-Yin Bridge and along the Shen-Hai Line along the Sutong Bridge Constitute, reflect the overall composition of the local characteristics.