基于空间网络的多阈值对象聚类

Multi-thresholds clustering objects on spatial network

为了提高空间网络上对象聚类的准确性,发现潜在的聚类结果,分析了多阈值选择的必要性,提取了空间网络上(以道路网为例)对象的相似性特征,首次提出了基于空间网络的支持向量回归多阈值方案,并将多阈值方案应用到已有的基于道路网络的对象聚类方法中,解决了已有聚类方法中聚类阈值选择困难的问题。性能分析及实验结果表明,多阈值对象聚类方案对真实的道路网络中的对象聚类是有效的。

To increase the accuracy of the clustering and find the hidden clusters on a spatial network, the necessity of multiple thresholds selection on a spatial network is analyzed, the similar feature of the objects is extracted, the scheme of multiple thresholds based on support vector regression（SVR） is proposed and the existing algorithm is improved.Performance analysis and experimental results indicate that the multiple thresholds scheme can achieve high efficiency and accuracy for clustering objects based in a road network.