北京测绘2025，Vol.39Issue(5)：730-735,6.DOI:10.19580/j.cnki.1007-3000.2025.05.025
基于地铁刷卡数据的嫌疑人识别方法
Suspect identification method based on subway card swipe data
摘要
Abstract
Passengers in urban public transportation systems have always been the main target for suspects.How to identify suspects based on passengers'travel behavior characteristics,in order to prevent theft incidents in public transportation,has become a key focus for security authorities.With the rapid development of big data,artificial intelligence,and other technologies,transit card swipe data has become an important foundation for mining passengers'movement patterns and conducting social awareness.However,existing methods for identifying anomalies from massive trajectory data still suffer from high false positive rates,making it difficult to effectively identify suspects.To address this,this paper proposed a two-step anomaly trajectory identification method.First,a portion of the samples were filtered through K-means clustering,and then the remaining samples were classified using One-Class support vector machine(SVM)for binary classification to identify suspects.Taking the subway card swipe data from Shanghai as an example,experimental results show that the two-step identification method of clustering followed by classification can effectively distinguish the abnormal trajectories of suspects.The suspects are mainly located near subway stations in downtown Shanghai.The results align with actual observations and can provide a reference for the further establishment of monitoring and tracking systems.关键词
公共交通/地铁刷卡/嫌疑人/异常轨迹识别/K-均值聚类/支持向量机(SVM)Key words
public transportation/subway card swipe/suspect/anomaly trajectory identification/K-means clustering/support vector machine(SVM)分类
天文与地球科学引用本文复制引用
郭思慧,李佳蓉,黄梦娜,刘高鸣..基于地铁刷卡数据的嫌疑人识别方法[J].北京测绘,2025,39(5):730-735,6.基金项目
北京市自然科学基金(4214069) （4214069）
