基于智能手机大数据的交通出行方式识别研究  

Travel mode recognition based on smart phone big data

作　　者：李喆 孙健[2,3] 倪训友 Li Zhe Sun Jian Ni Xunyou(Shanghai Municipal Transportation Design Institute Co. , Ltd. , Shanghai 200030, China State Key Laboratory of Ocean Engineering Transportation Research Center, Sclwol of Naval Architecture, Ocean ＆ Civil Engineering, Shanghai Jiao Tong University, Shanghai 200240, China)

机构地区：[1]上海市政交通设计研究院有限公司综合交通所,上海200030 [2]上海交通大学船舶海洋与建筑工程学院海洋工程国家重点实验室,上海200240 [3]上海交通大学船舶海洋与建筑工程学院交通研究中心,上海200240

出　　处：《计算机应用研究》2016年第12期3527-3529,3558,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(71101109);上海市"科技创新行动计划"软科学研究重点资助项目(15692105400)

摘　　要：智能手机时代所产生的大数据能够为交通研究者带来大量信息,基于智能手机采集交通出行大数据,再利用基于粒子群的支持向量机模型进行交通出行方式识别研究。在分析数据特点的基础上提出用于建模的特征变量,之后使用粒子群算法优化支持向量机参数,并基于成都市的实证数据进行模型的训练与出行方式识别研究。研究结果表明,该模型识别正确率为95.1%,高于决策树、BP神经网络、基于网格搜索的支持向量机模型,且该模型在时间效率方面具有明显的优越性,因而在出行方式识别方面具有良好的现实意义。The big data generated by smart phones can bring a lot of information for traffic investigators, this paper proposed a model based on particle swarm optimization and support vector machine to recognize different travel modes based on the smart-phone data. After analyzing the characteristics of data collected by smartphones, it proposed several feature variables for mode-ling. Further on, it used particle swarm optimization for optimizing the support vector machine model, which was trained and tested for travel mode recognition based on the empirical data in Chengdu, Sichuan province. The results indicate that, the recognition accuracy of the proposed model attains 95.1%, is higher than that of the decision trees, back propagation neural network model and the support vector machine based on grid search optimization. The time efficiency of the proposed model has good performance in urban travel mode recognition.

关 键 词：粒子群 支持向量机 出行方式识别 智能手机大数据 模式识别 

分 类 号：U116.2[交通运输工程]