Unsupervised Methods for a Personalised Route Recommendation System

There are many vehicle navigation systems in existence. They provide directions following routes based on speed, traffic, major or minor roads and places of interest. Although guidance systems are getting increasingly more sophisticated, to our knowledge none of them can provide a recommendation based on subjective preferences. This contribution studies classification algorithms that can provide this very feature. Given all possible paths between geographical locations, the proposed method classifies the available paths, making possible the choice of a trajectory with subjective characteristics.