A new approach for learning user preferences for a ridesharing application

Ridesharing has the potential to relieve some transportational issues such as traffic congestion, pollution and high travel costs. In this paper, we focus on the process of matching drivers and prospective riders more effectively, which is a crucial challenge in ridesharing. A novel approach is proposed in ride-matching which involves learning user preferences regarding the desirability of a choice of matching; this could then maintain high user satisfaction, thus encouraging repeat usage of the system. An SVM inspired method is developed which is able to learn a scoring function from a set of pairwise comparisons, and predicts the satisfaction degree of the user with respect to specific matches. To assess the proposed approach, we conducted some experiments on a commercial ridesharing data set. We compare the proposed approach with five rival strategies and methods, and the results clearly show the merits of our approach.