Conditional Preference Learning for Personalized and Context-Aware Journey Planning

Conditional preference networks (CP-nets) have recently emerged as a popular language capable of representing ordinal preference relations in a compact and structured manner. In the literature, CP-nets have been developed for modeling and reasoning in mainly toy-sized combinatorial problems, but rarely tested in real-world applications. Learning preferences expressed by passengers is an important topic in sustainable transportation and can be used to improve existing journey planning systems by providing personalized information to the passengers. Motivated by such needs, this paper studies the effect of using CP-nets in the context of personalized and context-aware journey planning. We present a case study where we learn to predict the journey choices by the passengers based on their historical choices in a multi-modal urban transportation network. The experimental results indicate the benefit of the conditional preference in passengers' modeling in context-aware journey planning.