计算机应用研究2017，Vol.34Issue(1)：136-140,5.DOI:10.3969/j.issn.1001-3695.2017.01.029
一种基于C FS FD P改进算法的重要地点识别方法研究
Research on important places identification method based on improved CFSFDP algorithm
摘要
Abstract
To solve the problem that CFSFDP clustering algorithm could not be applied to important places identification for the reason that it was unable to decide the cluster number with CFSFDP.This paper introduced a cluster center automatic choosing strategy to improve the algorithm.The strategy regarded the trends of cluster center weights changing as a rule with which decide the cluster center points automatically,avoiding the error brought by decision graph method.As a result,the method combing with CFSFDP algorithm,data preprocessing and reverse geocoding technology could improve the accuracy of important identification.The experiment chose Foursquare data as an example,the result shows that improved algorithm has higher accuracy rate and lower computation compared to DBSCAN.It also proves that the method has the advantage of other methods in handling the problem that important places identification with sparse location dataset.关键词
重要地点识别/速度剪枝/基于密度的聚类/密度峰值/簇中心Key words
important place identification/velocity pruning/density based clustering/density peak/cluster center分类
信息技术与安全科学引用本文复制引用
马春来,单洪,马涛,朱立新..一种基于C FS FD P改进算法的重要地点识别方法研究[J].计算机应用研究,2017,34(1):136-140,5.基金项目
国防重点实验室基金资助项目 （）
