基于近邻传播聚类的铁路客运节点类别划分

为简化旅客列车开行方案优化编制问题，采用聚类法对铁路客运节点进行类别划分。选取2014年日均旅客发送量排名前100位客运节点的相关属性数据为实例，以铁路客运节点的类别划分依据作为属性变量，首先运用分层聚类中的凝聚法对属性变量进行聚类，然后根据简化的客运节点变量指标，运用近邻传播算法对客运节点样本其进行聚类，并引用CH、KL、IGP等三种聚类有效性指标对聚类结果加以分析。研究结果表明，将100个客运节点分为5个类别时，具有最好的聚类效果，可为旅客列车开行方案的设计奠定基础。

Classification of Railway Passenger Transport Nodes Based on Affinity Propagation Cluster

To simplify the problem of optimizing passenger line plan, this paper adopted clustering method to classify railway passenger transport nodes. 100 top passenger transport nodes were selected in order of daily passenger dispatch volumes in 2014 as an example, and classifying foundation of passenger transport nodes was taken as property variables. Firstly, property variables were clustered by hierarchical clustering. And passenger transport nodes samples were clustered by affinity propagation algorithms according to the simplified nodes indexes. Finally, three clustering effectiveness indexes contained CH, KL and IGP indexes were analyzed to the clustering consequence. The result showed that it was of the best effect while the 100 passenger transport nodes mentioned above were divided into 5 sorts, and it laid a foundation for the design of passenger train plan.