Identifying and Analyzing the Prevalent Regions of a Co-Location Pattern Using Polygons Clustering Approach

Abstract

Given a co-location pattern consisting of spatial features, the prevalent region mining process identifies local areas in which these features are co-located with a high probability. Many approaches have been proposed for co-location mining due to its key role in public safety, social-economic development and environmental management. However, traditionally, most of the solutions focus on itemsets mining and results outputting in a textual format, which fail to adequately treat all the spatial nature of the underlying entities and processes. In this paper, we propose a new co-location analysis approach to find the prevalent regions of a pattern. The approach combines kernel density estimation and polygons clustering techniques to specifically consider the correlation, heterogeneity and contextual information existing within complex spatial interactions. A kernel density estimation surface is created for each feature and subsequently the generated multiple surfaces are combined into a final surface with cell attribute representing the pattern prevalence measure value. Polygons consisting of cells are then extracted according to the predefined threshold. Through adding appended environmental data to the polygons, an outcome of similar groups is achieved using polygons clustering approach. The effectiveness of our approach is evaluated using Points-of-Interest datasets in Shenzhen, China.

The Rural Development Policy in Extremadura (SW Spain): Spatial Location Analysis of Leader Projects

“…The value obtained represents the result of a calculation made on several points around it, so that the points farthest from the pixel center has a lower weight in the calculation. Thus, this method generates a smooth curved surface on each studied point using local neighborhood calculations defined by cell clusters or pixels and a radial symmetric function [42]. In this way, the highest density value is given over the point and tends to decrease progressively as it moves away from it until the bandwidth, where the density value reaches 0 [43].…”

High Influencing Pattern Discovery over Time Series Data

“…Generate random numbers for influential media: Based on synthetic datasets, one generates random numbers of influential media in [1,2] for influencing instances and generates random numbers of influential media in [10,15] for influenced instances; the generated numbers are randomly assigned to instances in a distributed manner as much as possible at five specific moments. The amount of influential media in influencing instances equals that of influential media in influenced instances at adjacent moments, and the total amount of influential media is specified for a specific synthetic dataset, e.g., five thousand influential media are generated for the Syn-1 dataset.…”

“…Construct new proximate relations: In order to construct new proximate relations between instances of distinct features, we reviewed the theoretical basis by which traditional spatial co-location pattern mining adopts spatial proximity, i.e., the first law of geography (or called Tobler's first law (TFL)), alleging that everything is related to everything else but near things are more related to each other [8,9]. Based on this law, traditional spatial co-location pattern mining uses spatial distance or combined distance [10] to measure spatial nearness. In epidemic dispersal scenarios, viruses spread with infectors among streams of people, so neighbor cities in a space may have no or little association when there are no or few personnel exchanges; meanwhile, a strong association may exist between two cities that are far away but have frequent exchanges on a large scale, and in these contexts, spatial distance is no longer the dominant factor for reflecting the sematic proximate relations between instances on influence.…”

Maximal Instance Algorithm for Fast Mining of Spatial Co-Location Patterns

“…A multi-level method was developed to identify regional co-location patterns in two steps: first, global co-location patterns were detected and other non-prevalent co-location patterns were identified as candidates for regional co-location patterns; second, an adaptive spatial clustering method was applied to detect the sub-regions where regional co-location patterns are prevalent [13]. Some researchers have studied the relationship between the spatial co-location pattern and clustering [14][15][16], which provides a new idea for co-location mining algorithm. Ouyang et al studied the co-location mining problem for fuzzy objects and proposed two new kinds of co-location pattern mining for fuzzy objects, single co-location pattern mining (SCP) and range co-location pattern mining (RCP), for mining co-location patterns at a membership threshold or within a membership range [17].…”