Personalized and Context-Aware Recommendation of Running Routes
Link zur Publikation:
Route recommendations are typically based on methods which generate an optimal sequence of road segments between two points. Regarding running routes, getting from A to B is less important than running on a route which meets specific requirements relevant to a runner. Map data is usually obtained for car routing problems. Data on running routes, particularly routes which are in remote areas, might only consist of GPS traces with no additional information such as terrain or lighting. In order to tackle these issues, running routes are considered as objects described with different attributes. As a result, recommendation of running routes becomes a multi-criteria decision making problem. We describe how this objects and their attributes can be obtained. We incorporate personalization by applying a location-based filtering mechanism to determine running routes in the vicinity of a user or a user-specified location. This selection is then sorted by an ontology supported ranking mechanism which evaluates route attributes in relation to a user profile, user context and collaborative features.