基于半监督SVM的交通方式特征分析和识别

Transportation Mode Feature Analysis and Recognition Based on Semi-supervised Support Vector Machine

为解决基于手机信令数据识别大规模用户交通方式问题,多维度分析出行方式特征,提出结合主动学习和Tri-training的半监督支持向量机算法。以手机信令出行链为基础,将出行特征划分为距离、时间、速度、出行者属性等四类,并进一步研究多维度特征及其计算方法 ,基于有向无环图设计一种结合主动学习与基于Tri-training的半监督多分类支持向量机。运用HY市手机信令数据构建样本集并训练该分类器,与多种监督学习分类算法进行比较。结果表明主动学习构造的富含信息的已标记样本集可以减少半监督学习的迭代次数,Tri-training半监督支持向量机可以通过大量未标记样本提升分类器准确率,结合主动学习与Tri-training半监督支持向量机算法可以有效地识别手机信令数据出行链的交通方式。

In order to solve the problem of recognizing transportation modes based on large numbers of mobile phone signaling data,multi-dimensional transportation characteristics are analyzed and Semisupervised support vector machine algorithm combined with active learning and Tri-training is proposed.the transportation characteristics are divided into four categories:distance,time,speed,and traveler attributes,and multi-dimensional features and their calculation methods are further studied based on mobile phone signaling travel chain.A semi-supervised multi-classification support vector machine combining active learning and Tri-training is studied based on directed acyclic graphs.The HY city mobile phone signaling data was used to construct a sample set and train the classifier to compare with the supervised learning classification algorithm.The results show that the active learning structure of the information-rich labeled sample set can reduce the number of iterations of semi-supervised learning.The Tri-training semi-supervised support vector machine can improve the accuracy of the classifier through a large number of unlabeled samples.Combined with the active learning and Tri-training semi-supervised support vector machine algorithm,the mobile location data transportation mode can be effectively discriminated.