Research on improved DBSCAN clustering algorithm based on track data
In order to study the simulation training track data clustering,this paper aims at the problem of in-accurate parameter selection and low clustering accuracy of the traditional DBSCAN algorithm to propose an im-proved DBSCAN clustering algorithm.Firstly,KNN algorithm is used to calculate the neighborhood radius and obtain the initializing core data object for DBSCAN clustering to realize rough clustering.Then,according to the characteristics of data objects,heading features are added for a secondary clustering,which not only solves the dif-ficulty of randomly initialized core point and parameter selection of DBSCAN algorithm,but also adds features that reflect the direction of the data.Finally,a simulation experiment is carried out.The experimental results show that the improved DBSCAN algorithm has better clustering effect than the traditional algorithm.
simulation trainingDBSCAN algorithmsecondary clusteringadaptive parameter selectiontrack data
国家科技期刊平台
NSTL
万方数据
