基于聚类分析的综合杆设计应用

Clustering Analysis and Its Application in Design of Multifunctional Integrated Poles

道路杆件具有数值和分类两种数据类型,应用聚类算法开展道路杆件数据的探索性分析。针对杆件坐标的数值型变量,采用基于欧式距离的K-means算法进行空间聚类,而对杆件类型、杆件设施类型及高度等分类型变量采用基于汉明距离的K-modes算法进行属性聚类。实例应用表明:K-means算法可将现状杆件样本沿道路划分为相应的簇,簇数可作为综合杆工程量的参照,而K-modes算法得出的分类属性聚类结果可作为综合杆件选型的依据。

The road poles have two data types:numerical data type and categorical data type.The clustering algorithm is used to carry out an exploratory analysis of road pole data.For the numerical variables of the coordinates of the poles,the K-means algorithm based on Euclidean distance is used for spatial clustering,and the K-modes algorithm based on Hamming distance is adopted to handle clustering algorithms with the categorical data type,such as pole’s types,facility types,and heights.The example application shows that the K-means algorithm can divide the existing pole samples into corresponding clusters along the road,and the number of clusters can be used as a reference for the amount of multifunctional integrated poles,the categorical attribute clustering results obtained by the K-modes algorithm can serve as a useful basis for the type selection of multifunctional integrated pole.