ISBN: 978-981-11-0008-6 DOI: 10.18178/wcse.2016.06.100
Urban Trip Requests Prediction: An Operator’s Perspective
Abstract— With the help of e-hailing apps, such as Uber, finding passengers is no longer a problem for
drivers. Transportation network companies’ operators are concerned about how to allocate their cars in the
urban area to maximize the company’s profit. To answer that question, they have to know how trip requests
are distributed over time and regions. In this paper, we present Gaussian Mixture Experts (GME), to partition
the city into multiple regions and estimate of the trip request density with fixed Gaussian components but
time-variant weights. We utilize the periodicity to predict the trip requests by context filtering. Our method is
tested on real industrial urban trip request data. The result shows that our method achieves a good
approximation to the real condition.
Index Terms— trip request prediction, spatial data mining, Gaussian mixture experts, EM algorithm
Yong Tian, Ningyuan Huang, Weidong Liu, Jiaxing Song
Department of Computer Science and Technology, Tsinghua University, CHINA
Cite: Yong Tian, Ningyuan Huang, Weidong Liu, Jiaxing Song, "Urban Trip Requests Prediction: An Operator’s Perspective," Proceedings of 2016 6th International Workshop on Computer Science and Engineering, pp. 585-589, Tokyo, 17-19 June, 2016.