Adaptive, Multi-criteria Recommendations for Location-Based Services

Location-based services have faced a development from being a hype to be used by a large user community at any place and time. However, only a few approaches exist, that take into account social interactions and learn from them in order to refine recommendations of points of interests accordingly. This paper analyzes the influence factors of mobile users for the choice of interests and derives an adaptable ranking function, that is capable of adjusting preferential weights on certain influence factors in order to learn from user behavior using ontology evolution. The Cool City Use Case demonstrates the application of the approach in a big city and shows how this adaptive learning can improve social recommendations of points of interests.