| With the development of information technology, human access and massive data storage capacity of the rapid increases in data mining production of knowledge in some context, spatial data mining and knowledge discovery is the main research contents. Spatial Data Mining and Knowledge Discovery technology, make space for an effective improvement of knowledge acquisition. Location problem in public facilities, if the use of mathematical language of its considerations to model spatial data mining and knowledge to which the city will be proposed location of public facilities and provide valuable reference for decision-making. This paper mainly studies the theory of spatial data mining techniques and methods, and clustering algorithm optimization decision-making in public facility location application.In the basic theory,spatial data mining and knowledge discovery based on theory are reviewed. Knowledge of the spatial clustering concept, significance was explained, describing the cluster and cluster-level segmentation algorithm. Introduced the K-means algorithm, K-medoids algorithm, Clara algorithm, Clarans algorithm and the algorithm and the algorithm flow of ideas, and the performance of various clustering algorithms and the pros and cons of clustering results were compared.Expansion of research in order to analyze the object of public facility location, in the traditional clustering algorithm application in this regard based on the full account of the barrier factors and differences in road traffic conditions of the impact of factors on the cluster results, the algorithm improvements to optimize the clustering results. At the same time with the K-means algorithm and the effective combination of simulated annealing, to overcome the traditional K-means algorithm and the traditional K-mediods algorithm of the state. And traditional solutions by contrast, reflects the operation of the proposed method and results of time-effectiveness.In experimental studies, based on the improved algorithm, using C + + programming language to achieve its encoded and carried out in the Visual Studio 2005 compile and run. And to a new city, for example, access to residential use Google earth coordinates as input data, respectively, this improved algorithm and application of traditional clustering algorithm, and then apply a graph depicting point MapInfo experimental results were compared.In this paper, the clustering of spatial data mining applications in public facilities to conduct a preliminary study and design of the algorithm based on coded implementation. Spatial data mining to knowledge on the application of construction of public facility location decision applied the system to explore. |
