Social Sensing for Improving the User Experience in Orienteering

We illustrate an application of social sensing for automatically customizing the solutions to orienteering problems with category constraints. The aim is to improve the user satisfaction in tourist trip planning provided by a previously developed efficient and effective algorithm. The system tries to perceive the possible user interests by sensing their social network profile, induced by computing topics arising from their posts. A set of semantic distance between the user profile and the descriptions of the points of interest (POIs) is then computed, giving an ordered list of POIs according to the user interests. The orienteering algorithm is then run taking into account this list of POIs sorted by the semantic closeness to the user posts in social networks. A preliminary running example has been setup and the results are promising in terms of both accuracy and performance.