Adaptive Discovery of Multi-level Spatial Co-location Patterns
Adaptive Discovery of Multi-level Spatial Co-location Patterns
By Jiannan Cai, Apr. 10, 2019
Embedded Files
Background
Background
Because of spatial heterogeneity, spatial co-location patterns are usually geographically regional. Regional spatial co-location patterns can be represented as a collection of spatial features that are frequently located together in certain localities (i.e. sub-regions) in the study area. The discovery of regional co-location patterns can facilitate the understanding of the spatial dependency of different spatial features at the micro-scales. However, regional co-location patterns remain challenging to discover for two main reasons : 1) sub-regions with co-located features are unknown a priori, and 2) the densities of the instances of a regional co-location pattern may vary across a study area. Thus, this study develops a multi-level method, in which the discovery of regional co-location patterns is modeled as a special clustering problem.
Innovations
Innovations
Spatial co-location patterns that are not prevalent at global level are first identified as candidates for regional co-location patterns, and then an adaptive pattern spatial clustering method is developed to detect the underlying sub-regions of each candidate regional pattern.
Figure 1. Discovery of candidate sub-regions based on adaptive pattern clustering method. (Deng et al., 2017)
To improve computational efficiency, an overlap method was developed to deduce the sub-regions of (k+1)-size co-location patterns from the sub-regions of k-size co-location patterns.
Figure 2. The overlap relationship between sub-regions of a 3-size pattern and its 2-size sub-patterns. (Deng et al., 2017)
Applications
Applications
The proposed method were used to discover the symbiosis among five plant species in a wetland located in northeast China with a dataset recording the locations of five plant species. The discovered global and regional co-location patterns can facilitate domain-expert explorations of relationships between plant species and environmental factors.
Figure 3. Detecting regional co-locations from the wetland species dataset in Northeast China. (Deng et al., 2017)
Table 1. The global and regional co-locations detected from the wetland species dataset in Northeast China. (Deng et al., 2017)
Table 1. The global and regional co-locations detected from the wetland species dataset in Northeast China. (Deng et al., 2017)
References
References
Jiannan Cai, Qiliang Liu, Min Deng, Jianbo Tang and Zhanjun He., 2018. Adaptive detection of statistically significant regional spatial co-location patterns. Computers, Environment and Urban Systems, 68, 53-63. (Link)
Min Deng, Jiannan Cai, Qiliang Liu, Zhanjun He and Jianbo Tang., 2017. Multi-level method for discovery of regional co-location patterns. International Journal of Geographical Information Science, 31(9), 1846-1870. (Link)
Google Sites
Report abuse