Earticle

초록

영어

In line with the development of platform business, O2O (OnlinetoOffline) platform business have emerged in this era. One of the representative O2O platform business is ridehailing service platform. However, the potential inefficiency has been reported in several news. To deal with this potential inefficiency, demand forecasting is conducted in this paper. Timevarying Poisson process and weighted timevarying Poisson process are employed to model the geospatial count data. In addition, several matrix completion techniques also adapted to preprocess the data. Weighted timevarying Poisson process outperforms than timevarying poisson process. In further studies, deep learning models and experiments are conducted to predict future demands, and optimization problem regarding resource allocation will be discussed.

목차

Abstract
Introduction
Methods
Result
Discussion
Future Work
References

저자

• Sumin Lim [ KAIST ]
• Chul Ho Lee [ KAIST ]
• Yasin Ceran [ KAIST ]
• Young U. Ryu [ University of Texas at Dallas ]

참고문헌