Estimating Passenger Preferences Using Implicit Relevance Feedback for Personalized Journey Planning

Personalized journey planning is becoming increasingly popular, due to strong practical interests in high-quality route solutions aligned with commuter preferences. In a journey planning system, travelers are not just mere users of the systems, instead they represent an active component willing to take different routes based on their own preferences, e.g., the fastest, least number of changes, or cheapest journey. In this work, we propose a novel preference estimation method that incorporates implicit relevance feedback methods into the journey planner, aiming to provide more relevant journeys to the commuters. Our method utilizes commuters’ travel history to estimate the corresponding preference model. The model is adaptive and can be updated iteratively during the user/planner interactions. By conducting experiments on a real dataset, it can be demonstrated that the proposed method provide more relevant journeys even in absence of explicit ratings from the users.