作　　者： ; ; ; ;

机构地区： 中山大学地理科学与规划学院

出　　处： 《热带地理》 2012年第2期179-184,共6页

摘　　要： 城镇地价的空间分异对其动态监测片区划定具有重要的参考作用,而空间聚类则是对该问题进行分析的一种有效方法。文中在分析空间聚类本质的基础上,设计了一种基于样本数据的粒子群双重空间聚类挖掘模型,并以北海市二手房地价空间分布挖掘为实例进行了应用。结果表明:粒子群双重空间聚类在空间聚类效果上优于传统K-均值聚类法,尤其是对于任意形状聚类簇的挖掘;然而,由于标准粒子群优化算法机理的限制,其聚类效果随聚类中心的增多而呈下降趋势。总体而言,该方法能利用有限的样本挖掘质量较好的城镇地价空间分异区,可为城镇地价动态监管提供相关决策基础。 The spatial distribution of urban land price takes an important role in dividing of land for its price dynamic monitoring, and the spatial clustering is an effective method in analysis of that issue. In this paper a dual spatial clustering model with particle swarm optimization is designed according to the essence of spatial clustering, and the model is applied to the mining of spatial distribution of second-hand housing price in Beihai, Guangxi. Results demonstrate that this method is better than the traditional K-Means spatial clustering, especially for the mining of arbitrary shape clustering. However, its clustering ability is dropping with the increase of the clustering center because of the limitation of algorithm itself. Overall, the model can mine the spatial distribution of urban land price with good quality, so it can provide decision knowledge for urban land price dynamic suoervision.

关 键 词： 空间聚类 粒子群 城镇地价

领　　域： [天文地球] [天文地球]