Tensor-based User Trajectory Mining

被引：0

作者：

Yu, Chen ^{[1
]}

Hong, Qinmin ^{[1
]}

Yao, Dezhong ^{[1
]}

Jin, Hai ^{[1
]}

机构：

[1] Huazhong Univ Sci & Technol, Sch Comp Sci & Technol, Serv Comp Technol & Syst Lab, Cluster & Grid Comp Lab,Big Data Technol & Syst L, Wuhan 430074, Hubei, Peoples R China

来源：

COMPUTER SYSTEMS SCIENCE AND ENGINEERING | 2018年 / 33卷 / 02期

关键词：

Data mining; hot route discovery; GPS logs; route recommendation; MOBILE; DECOMPOSITIONS;

D O I：

暂无

中图分类号：

TP3 [计算技术、计算机技术];

学科分类号：

0812 ;

摘要：

The rapid expansion of GPS-embedded devices has showed the emerging new look of location-based services, enabling such offerings as travel guide services and location-based social networks. One consequence is the accumulation of a rich supply of GPS trajectories, indicating individuals' historical position. Based on these data, we aim to mine the hot route by using a collaborative tensor calculation method. We present an efficient trajectory data processing model for mining the hot route. In this paper, we rst model the individual's trajectory log, extract sources and destinations, use map matching to get the corresponding road segments, and nally apply the source-destination-road segments tensor in order to compute the recommended hot route. To prove the validity and efficiency of the method, we conduct a collaborative route recommendation system, and the experimental result indicated that the solution can recommend a route with considerable accuracy

引用

页码：87 / 94

页数：8