自组织双重空间聚类算法的城市扩张结构分析应用
|
摘要
双重空间聚类是能顾及空间连续性和属性相似性的空间数据分析,而常规空间聚类算法难以同时顾及2方面的约束条件。本文采用自组织双重空间聚类算法,对城市扩张结构分析进行了研究。通过改造自组织特征映射的最佳匹配神经元搜索的算法机制,在空间域和属性域进行迭代聚类搜索,实现了自组织双重空间聚类。以武汉市扩张斑块的位置信息和扩张程度指数为输入数据,使用自组织双重空间聚类算法,实现了城市扩张动态结构的识别。自组织双重空间聚类算法使得聚类结果,既在空间域上连续,又在属性域上相近,算法过程具有自组织性,减少了人为影响。
Dual spatial clustering is an exploratory data analysis that deals with spatial contiguity and attributive similarity. Conventional spatial clustering methods cannot perform effective clustering in spatial and attribute domains simultaneously. This study employs SOFM(Self-Organizing Feature Mapping) to solve dual spatial clustering problems, and then verify the proposed method in the analysis of urban expansion structure. By modifying the algorithm of best matching neuron searching in SOFM, we manage to perform clustering in both spatial and attribute domains. The algorithm includes two independent self-organizing clustering processes. The first one includes a spatial constraint, and the other one includes an attribute constraint. The final result is generated by merging the corresponding two results that derived separately from the two processes. The analysis of the structure of urban expansion of Wuhan city is used as a case study. We feed the proposed model with the location information and the expansion degree information of newly grown urban patches, and the generated dual clustering results could clearly illustrate the spatial structure of urban expansion. As a conclusion, the self-organizing dual spatial clustering method can generate spatial continuous and attributive similar clusters with little artificial interference.
Dual spatial clustering is an exploratory data analysis that deals with spatial contiguity and attributive similarity. Conventional spatial clustering methods cannot perform effective clustering in spatial and attribute domains simultaneously. This study employs SOFM(Self-Organizing Feature Mapping) to solve dual spatial clustering problems, and then verify the proposed method in the analysis of urban expansion structure. By modifying the algorithm of best matching neuron searching in SOFM, we manage to perform clustering in both spatial and attribute domains. The algorithm includes two independent self-organizing clustering processes. The first one includes a spatial constraint, and the other one includes an attribute constraint. The final result is generated by merging the corresponding two results that derived separately from the two processes. The analysis of the structure of urban expansion of Wuhan city is used as a case study. We feed the proposed model with the location information and the expansion degree information of newly grown urban patches, and the generated dual clustering results could clearly illustrate the spatial structure of urban expansion. As a conclusion, the self-organizing dual spatial clustering method can generate spatial continuous and attributive similar clusters with little artificial interference.
引文
[1]Lin C R,Liu K H,Chen M S.Dual clustering:Integrating data clustering over optimization and constraint domains[J].IEEE Trans.Knowledge and Data Engineering,2005,17(5):628-637.
[2]Jiao L M,Liu Y L.Research on self-organizing clustering of spatial points[C].Proceedings of SPIE Vol.6751,the15th International Conference on Geoinformatics,Nanjing,2007:5.
[3]Chawla S,Shekhar S.Modeling spatial dependencies for mining geospatial data:An introduction[A].In:Miller H J,Han J W(eds.).Geographic Data Mining and Knowledge discovery(GKD)[M].Boca Raton,FL:Taylor and Francis,2001.
[4]Han J,Kamber M,Tung A K H.Spatial clustering methods in data mining:A survey[A].In:H.Miller H J,Han J W(eds.).Geographic Data Mining and Knowledge Discovery[M].Boca Raton,FL:Taylor and Francis,2001.
[5]Halkidi M,Batistakis Y,Vazirgiannis M.On Clustering Validation Techniques[J].Journal of Intelligent Information Systems,2001,17(2-3):107-145.
[6]Tai C H,Dai B R,Chen M S.Incremental clustering in geography and optimization spaces[C].Proceedings of PAKDD’07,2007:272-283.
[7]Zhang B,Yin W J,Xie M,et al.Geo-spatial clustering with non-spatial attributes and geographic non-overlapping constraint:A penalized spatial distance measure[C].Proceedings of PAKDD'07,2007:1072-1079.
[8]Wang X,Hamilton H J.DBRS:A density-based spatial clustering method with random sampling[C].Proceedings of the 7th PAKDD,Seoul,Korea,2003:563-575.
[9]Zhou J G,Guan J H,Li P X.DCAD:A dual clustering algorithm for distributed spatial databases[J].Geo-spatial Information Science,2007,10(2):137-144.
[10]Henriques R,Bacao F,Lobo V.Exploratory geospatial data analysis using the Geo SOM suite[J].Computers,Environment and Urban Systems,2012,36:218-232.
[11]李光强,邓敏,程涛.一种基于双重距离的空间聚类方法[J].测绘学报,2008,37(4):484-488.
[12]周翠竹,朱建军,石岩.一种基于双重距离约束的多层次空间聚类方法[J].测绘科学,2014,39(10):98-101.
[13]刘启亮,邓敏,石岩,等.一种基于多约束的空间聚类方法[J].测绘学报,2011,40(4):509-516.
[14]Cao F,Ge Y,Wang J F.Optimal discretization for geographical detectors-based risk assessment[J].GIScience&Remote Sensing,2013,50(1):78-92.
[15]Wang J F,Li X H,Christakos G,et al.Geographical detectors-based health risk assessment and its application in the neural tube defects study of the Heshun Region,China[J].International Journal of Geographical Information Science,2010,24(1):107-127.
[16]Jiao L M,Liu Y L,Zou B.Self-organizing dual clustering considering spatial analysis and hybrid distance measures[J].Science in China-Earth Science,2011,54(8):1268-1278.
[17]Liu X P,Li X,Chen Y M,et al.A new landscape index for quantifying urban expansion using multi-temporal remotely sensed data[J].Landscape Ecology,2010,25:671-682.
[2]Jiao L M,Liu Y L.Research on self-organizing clustering of spatial points[C].Proceedings of SPIE Vol.6751,the15th International Conference on Geoinformatics,Nanjing,2007:5.
[3]Chawla S,Shekhar S.Modeling spatial dependencies for mining geospatial data:An introduction[A].In:Miller H J,Han J W(eds.).Geographic Data Mining and Knowledge discovery(GKD)[M].Boca Raton,FL:Taylor and Francis,2001.
[4]Han J,Kamber M,Tung A K H.Spatial clustering methods in data mining:A survey[A].In:H.Miller H J,Han J W(eds.).Geographic Data Mining and Knowledge Discovery[M].Boca Raton,FL:Taylor and Francis,2001.
[5]Halkidi M,Batistakis Y,Vazirgiannis M.On Clustering Validation Techniques[J].Journal of Intelligent Information Systems,2001,17(2-3):107-145.
[6]Tai C H,Dai B R,Chen M S.Incremental clustering in geography and optimization spaces[C].Proceedings of PAKDD’07,2007:272-283.
[7]Zhang B,Yin W J,Xie M,et al.Geo-spatial clustering with non-spatial attributes and geographic non-overlapping constraint:A penalized spatial distance measure[C].Proceedings of PAKDD'07,2007:1072-1079.
[8]Wang X,Hamilton H J.DBRS:A density-based spatial clustering method with random sampling[C].Proceedings of the 7th PAKDD,Seoul,Korea,2003:563-575.
[9]Zhou J G,Guan J H,Li P X.DCAD:A dual clustering algorithm for distributed spatial databases[J].Geo-spatial Information Science,2007,10(2):137-144.
[10]Henriques R,Bacao F,Lobo V.Exploratory geospatial data analysis using the Geo SOM suite[J].Computers,Environment and Urban Systems,2012,36:218-232.
[11]李光强,邓敏,程涛.一种基于双重距离的空间聚类方法[J].测绘学报,2008,37(4):484-488.
[12]周翠竹,朱建军,石岩.一种基于双重距离约束的多层次空间聚类方法[J].测绘科学,2014,39(10):98-101.
[13]刘启亮,邓敏,石岩,等.一种基于多约束的空间聚类方法[J].测绘学报,2011,40(4):509-516.
[14]Cao F,Ge Y,Wang J F.Optimal discretization for geographical detectors-based risk assessment[J].GIScience&Remote Sensing,2013,50(1):78-92.
[15]Wang J F,Li X H,Christakos G,et al.Geographical detectors-based health risk assessment and its application in the neural tube defects study of the Heshun Region,China[J].International Journal of Geographical Information Science,2010,24(1):107-127.
[16]Jiao L M,Liu Y L,Zou B.Self-organizing dual clustering considering spatial analysis and hybrid distance measures[J].Science in China-Earth Science,2011,54(8):1268-1278.
[17]Liu X P,Li X,Chen Y M,et al.A new landscape index for quantifying urban expansion using multi-temporal remotely sensed data[J].Landscape Ecology,2010,25:671-682.