Taxi call prediction for online taxicab platforms

Abstract

The online taxi-calling services have gained great popularity in the era of sharing economy. Comparing with the traditional taxi service, the online taxi-calling service is much more convenient and flexible for passengers, because the taxi platform can provide detailed travel arrangements, transparent estimated price and flexible means of calling in advance. To understand the call willingness of passengers and to increase the quantity of orders, it is important to predict whether a request will be converted to a call order. In this paper, we study the problem of taxi call prediction, which is one of the important components of taxi-calling service system. To solve the problem, we propose a prediction framework, which uses some classification models with combining the basic features such as the order and meteorology information and the personlized historical profiles. Extensive experimental evaluations on real taxi record data from an online taxi-calling service platform demonstrated the effectiveness of our approach.