A Simple but Quantifiable Approach to Dynamic Price Prediction in Ride-on-demand Services Leveraging Multi-source Urban Data
Suiming Guo,Chao Chen,Jingyuan Wang,Yaxiao Liu,Ke Xu,Daqing Zhang,Dah Ming Chiu
DOI: https://doi.org/10.1145/3264922
2018-01-01
Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies
Abstract:Ride-on-demand (RoD) services such as Uber and Didi are becoming increasingly popular, and in these services dynamic prices play an important role in balancing the supply and demand to benefit both drivers and passengers. However, dynamic prices also create concerns. For passengers, the "unpredictable" prices sometimes prevent them from making quick decisions: one may wonder if it is possible to get a lower price if s/he chooses to wait a while. It is necessary to provide more information to them, and predicting the dynamic prices is a possible solution. For the transportation industry and policy makers, there are also concerns about the relationship between RoD services and their more traditional counterparts such as metro, bus, and taxi: whether they affect each other and how.
In this paper we tackle these two concerns by predicting the dynamic prices using multi-source urban data. Price prediction could help passengers understand whether they could get a lower price in neighboring locations or within a short time, thus alleviating their concerns. The prediction is based on urban data from multiple sources, including the RoD service itself, taxi service, public transportation, weather, the map of a city, etc. We train a simple linear regression model with high-dimensional composite features to perform the prediction. By combining simple basic features into composite features, we compensate for the loss of expressiveness in a linear model due to the lack of non-linearity. Additionally, the use of multi-source data and a linear model enables us to quantify and explain the relationship between multiple means of transportation by examining the weights of different features in the model. Our hope is that the study not only serves as an accurate prediction to make passengers more satisfied, but also sheds light on the concern about the relationship between different means of transportation for either the industry or policy makers.