地理大数据多元流协同计算

项目来源

国(略)研(略)((略)D(略)

项目主持人

裴(略)

项目受资助机构

中(略)院(略)学(略)研(略)

立项年度

2(略)

立项时间

未(略)

项目编号

2(略)YFB0503604

项目级别

国(略)

研究期限

未(略) (略)

受资助金额

0(略)万(略)

学科

地(略)与(略)

学科代码

未(略)

基金类别

“地(略)与(略)”重点专项

地(略)流(略)时(略)关(略)群(略)识(略) (略)测(略)动(略)分(略) (略)件(略);(略)件(略);(略)o(略)p(略) (略)t(略)t(略)m(略)s(略)i(略)m(略)a(略)u(略)o(略)l(略)o(略) (略)s(略)i(略)p(略)e(略)r(略)g(略)i(略);(略)u(略)r(略)t(略)i(略);(略)n(略)c(略)r(略)u(略)a(略)y(略) (略)c(略)i(略)e(略)t(略)d(略)n(略) (略)e(略)e(略)t(略)m(略)t(略)

参与者

秦(略)朝(略)江(略)霞

参与机构

武(略)

项目标书摘要:地理(略)识别,是地理大数据(略)术。构建了地理多元(略)理论框架,实现了地(略)分析模型。(1)地(略)技术。构建了基于元(略)据表达模型,提出了(略) I指数的时空自相(略)多元流数据的时空演(略)据群聚模式识别技术(略)化对轨迹进行分割;(略)流速度以推算路段拥(略)踪识别出拥堵转向,(略)式。(3)地理流网(略)用格网划分对城市空(略)对轨迹进行出行OD(略)常出行活动的特征;(略),识别空间组团结构(略)结构分析技术。基于(略)交互网络,利用复杂(略)分析国家之间的交互(略)交互关系的变化趋势(略)流城市活动事件建模(略)轨迹数据和赛博空间(略)城市活动事件的事前(略)建模。(6)地理流(略)出租车、气象和空气(略)需求量和对应的下车(略)区域之间流量变化与(略)下居民出行模式和司(略)分析。

Applicati(略): Measure(略)entificat(略)raphic mu(略)am networ(略) is the k(略)gy of mul(略)ollaborat(略)ng of geo(略) data.We (略) the fund(略)oretical (略)f measure(略)entificat(略)raphic mu(略)am networ(略),realized(略)of patter(略)d analysi(略)phic mult(略) networks(略)of spatio(略)tocorrela(略)is of geo(略)ti-stream(略)sentation(略)eographic(略)am data b(略)lular aut(略)construct(略)temporal (略)tion anal(略) improvin(略) index is(略)nd spatio(略)olution c(略)ics of ge(略)lti-strea(略)discovere(略) of clust(略)recogniti(略)graphic s(略)The traje(略)are segme(略)on reside(略)d the cha(略)city.The (略)stributio(略)traffic f(略)y are cal(略)estimate (略)ngestion (略)reshold;t(略)racking i(略)dentify t(略)on steeri(略)congestio(略)of differ(略)g are ana(略)thod of c(略)cture det(略) geograph(略)etwork.Th(略)ce is div(略)asic proc(略)s based o(略)sion,the (略)airs are (略)rom the t(略)ata;the c(略)ics of da(略)activitie(略)ban resid(略)alyzed.Hi(略)clusters (略) units is(略)cognize t(略)cluster s(略)4)Method (略)structure(略)f geograp(略)network.N(略)eraction (略)e constru(略)on GDELT.(略)eristics (略) are expl(略)on the th(略)thod of c(略)ork,the i(略)relations(略)ations is(略)nd the ch(略)d and dev(略)e of the (略) relation(略)alyzed fr(略)ct of tim(略) of even (略)om city a(略)f geograp(略)The real (略)ic trajec(略)nd the so(略)data sets(略)pace are (略)the long-(略)s models (略)events,du(略)ents,and (略)vents are(略)ethod of (略)and predi(略)ographic (略)ts.Correl(略)sis to ob(略)lationshi(略)low chang(略)ions and (略) spatiote(略)ysis of r(略)avel mode(略)rs’route (略)odes are (略).

项目受资助省

北(略)

  • 排序方式:
  • 1
  • /
  • 1.地理多元流网络结构测度与识别(Measurement and Identification of Geographic Multiple Stream Network Structure)

    • 关键词:
    • 地理多元流、时空自相关、群聚模式识别、组团探测、动态结构分析、活动事件建模、流事件模拟、geographic multi-stream、spatiotemporal autocorrelation、clustering pattern recognition、cluster detection、dynamic structure analysis、activity event modeling、stream event simulation
    • 秦昆;康朝贵;陈江平;张霞;
    • 《武汉大学;武汉大学;武汉大学;武汉大学;》
    • 2019年
    • 报告

    地理多元流网络的结构测度与识别,是地理大数据多元流协同计算的关键技术。构建了地理多元流网络结构测度与识别的理论框架,实现了地理多元流网络模式挖掘与分析模型。(1)地理流数据时空自相关分析技术。构建了基于元胞自动机的地理多元流数据表达模型,提出了一种改进Moran’s I指数的时空自相关分析方法,揭示了地理多元流数据的时空演变特征。(2)地理流数据群聚模式识别技术。采用停留时间和速度变化对轨迹进行分割;计算速度分布及自由交通流速度以推算路段拥堵速度阈值;采用轨迹追踪识别出拥堵转向,并分析不同转向的拥堵模式。(3)地理流网络结构组团探测技术。采用格网划分对城市空间分割成基本处理单元,对轨迹进行出行OD对提取;分析城市居民日常出行活动的特征;对网格单元进行层次聚类,识别空间组团结构。(4)地理流网络动态结构分析技术。基于GDELT数据构建国家交互网络,利用复杂网络方法探索网络特征,分析国家之间的交互关系,并从时间角度探索交互关系的变化趋势和发展规律。(5)地理流城市活动事件建模技术。融合现实空间交通轨迹数据和赛博空间社交媒体数据集,实现对城市活动事件的事前、事中和事后长时过程的建模。(6)地理流事件模拟预测技术。基于出租车、气象和空气质量等数据,根据出租车需求量和对应的下车区域,进行关联分析得到区域之间流量变化与天气的关系,对不同天气下居民出行模式和司机路径选择模式进行时空分析。 Measurement and identification of geographic multiple stream network structure is the key technology of multi-stream collaborative computing of geographic big data.We constructed the fundamental theoretical framework of measurement and identification of geographic multiple stream network structure,realized the model of pattern mining and analysis of geographic multiple stream networks. (1)Method of spatiotemporal autocorrelation analysis of geographic multi-stream data. A presentation model of geographic multi-stream data based on cellular automation is constructed,a spatiotemporal autocorrelation analysis method improving Moran’s I index is proposed,and spatiotemporal evolution characteristics of geographic multi-stream data are discovered. (2)Method of cluster pattern recognition from geographic stream data. The trajectory data are segmented based on residence time and the change of velocity.The velocity distribution and free traffic flow velocity are calculated to estimate the road congestion velocity threshold;trajectory tracking is used to identify the congestion steering,and the congestion patterns of different steering are analyzed. (3)Method of cluster structure detection from geographic stream network. The urban space is divided into basic processing units based on grid division,the travel OD pairs are extracted from the trajectory data;the characteristics of daily travel activities of the urban residents are analyzed.Hierarchical clusters of the grid units is made to recognize the spatial cluster structures. (4)Method of dynamic structure analysis of geographic stream network. National interaction networks are constructed based on GDELT.The characteristics of networks are explored based on the theory and method of complex network,the interaction relationship among nations is analyzed,and the changing trend and developing rule of the interaction relationship are analyzed from the aspect of time. (5)Method of even modeling from city activities of geographic stream. The real space traffic trajectory data and the social media data sets in cyber space are integrated,the long-time process models before the events,during the events,and after the events are built. (6)Method of simulation and prediction of geographic stream events. Correlation analysis to obtain the relationship between flow change among regions and weather,and spatiotemporal analysis of residents’travel modes and drivers’route selection modes are carried out.

    ...
  • 排序方式:
  • 1
  • /