密度峰值聚类算法在管廊大数据挖掘中应用  

Application of Density Peak Clustering Algorithm in Big Data Mining of Underground Utility Tunnel

作　　者：马福军[1] 胡力勤 Ma Fujun;Hu Liqin(School of Building Equipmen,Zhejiang College of Construction,Hangzhou 311231,China)

机构地区：[1]浙江建设职业技术学院建筑设备学院,杭州311231

出　　处：《机电工程技术》2022年第2期94-97,共4页Mechanical & Electrical Engineering Technology

基　　金：浙江省自然科学基金研究项目(编号:LGF18F030005);2019年浙江省教育厅高校国内访问学者访问工程师项目(编号:FG2019114)。

摘　　要：为了准确、实时发现地下综合管廊运行和维护中的风险,将密度峰值聚类算法分析应用到地下综合管廊异常数据挖掘。密度峰值聚类算法分3个环节,离群数据的取舍、聚类中心的确定和以Voronoi图单元为基础的数据映射分配。通过实验分析,成功实现地下综合管廊环境中氧气浓度的大数据聚类,并得到不同氧气浓度数据聚类簇图像,直观地观察到氧气浓度的数据状态,通过该算法得到的数据簇聚类效果非常具有工程实际意义,能准确、实时预测管廊风险。In order to find the risk of the operation and maintenance of underground comprehensive pipe gallery accurately and in real time,the clustering algorithm of density peak to the mining of abnormal data of underground comprehensive pipe gallery was applied.The algorithm of density peak clustering was divided into three parts:the choice of outlier data,the determination of cluster center and the data mapping allocation based on Voronoi graph unit.Through experimental analysis,the large data cluster of oxygen concentration in underground comprehensive pipe gallery environment was successfully realized,and the cluster images of different oxygen concentration data were obtained.The data cluster status of oxygen concentration was observed intuitively.The clustering effect of data cluster obtained by the algorithm is very practical and can accurately and real-time predict the risk of the pipe gallery.

关 键 词：密度峰值聚类 地下综合管廊 大数据挖掘 

分 类 号：TP311.1[自动化与计算机技术—计算机软件与理论]