通信与信息技术Issue(3)：114-121,8.
5G热点区域聚类凸包算法检测方法研究
A 5G hotspot detection method based on an improved DBSCAN incremental convex hull algorithm
摘要
Abstract
With the increasing risk of public emergencies in society,enhancing the ability to prevent and detect emergencies has be-come an urgent issue to address.From the perspective of communication networks,an improved DBSCAN incremental convex hull algo-rithm for 5G hotspot detection has been proposed.Firstly,the latitude and longitude distribution data of 5G base stations are collected,and the Haversine Distance algorithm is used to transform the spatial distribution data into spherical matrix data,which serves as the in-put for clustering.Secondly,based on the clustering scenarios of 5G base station hotspots,the optimal clustering radius(Eps)and the min-imum number of aggregate points(MinPts)are selected as hyperparameters for the DBSCAN mining algorithm,dividing the matrix data in-to multiple clustered hotspots.Then,according to the distribution shape of the clusters,the Convex Hull algorithm is utilized to construct the shape of the hotspot areas,forming convex planar polygons.Finally,to further enhance the efficiency of hotspot clustering,a new busi-ness density assessment method is established,which is tailored to the actual application scenarios of business hotspots.By evaluating the clustered hotspot areas,hotspots with low clustering density are eliminated.The results show that this method can effectively detect dy-namically changing hotspot business areas in communication networks,capture emergencies in real-time,and is of great significance for enhancing the network's ability to prevent emergencies.关键词
DBSCAN聚类/半正矢距离/热点检测/增量凸壳算法/业务密度/突发事件Key words
DBSCAN/Haversine distance/Hotspots detection/Convex hull/Service density/Emergency分类
信息技术与安全科学引用本文复制引用
邓翠艳,齐小刚,姚旭清..5G热点区域聚类凸包算法检测方法研究[J].通信与信息技术,2025,(3):114-121,8.基金项目
2023年度山西省高等学校科技创新项目"5G网络智能精准节电研究"(项目编号:2023L517) （项目编号:2023L517）
