基于自适应网格密度算法的出行模式时空分析  

Spatiotemporal Analysis of Travel Modes Based on an Adaptive Grid Density Algorithm

作　　者：熊敏 石超峰 张玺[2] XIONG Min;SHI Chaofeng;ZHANG Xi(School of Economics and Management,Chongqing Jiaotong University,Chongqing 400074;School of Transportation,Chongqing Jiaotong University,Chongqing 400074)

机构地区：[1]重庆交通大学经济与管理学院,重庆400074 [2]重庆交通大学交通运输学院,重庆400074

出　　处：《计算机与数字工程》2024年第3期653-658,676,共7页Computer & Digital Engineering

基　　金：国家社会科学基金项目(编号:16BJL121);重庆市教委科学技术研究项目(编号:KJ1705148)资助。

摘　　要：针对出租车GPS原始轨迹数据中噪声和传统密度算法对大数据处理成本高,参数选择困难,易影响聚类效果等缺陷,给出了轨迹数据预处理方法和提出了一种基于自适应网格密度大数据区域挖掘算法。研究结果表明,自适应网格密度算法,能够有效地避免参数调节环节,样本空间适应性强,聚类质量高,与通常密度聚类算法相比,计算量小,计算效率高,给出的重庆市居民出行方式时空特征,符合实际,有应用价值。A trajectory data preprocessing method and a region mining algorithm based on adaptive grid density big data are proposed to address the drawbacks of noise in taxi GPS raw trajectory data,high processing costs for big data,difficulty in parame-ter selection,and susceptibility to clustering effects in traditional density algorithms.The research results indicate that the adaptive grid density algorithm can effectively avoid the parameter adjustment process,has strong sample space adaptability,and high clus-tering quality.Compared with conventional density clustering algorithms,it has lower computational complexity and higher computa-tional efficiency.The spatiotemporal characteristics of Chongqing residents'travel patterns provided are in line with reality and have practical value.

关 键 词：出租车GPS轨迹 数据清理 网格密度 热点区域 聚类挖掘 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]