Recommender system can be used to benefit travellers who are on the lookout of places to visit. Most of the times there are only few major places which in general everyone would know for planning visits but in between all these there are some serene unknown parts which people miss because they never knew about them. This is because of the availability of overwhelming and varied amount of information related to travel.
While designing a recommender system the fact that information is overwhelming and varied has to be taken into account. Most of the traditional travel recommender systems are simple content based recommenders which at each stage of the traveller’s interaction decide which product(s) to recommend to the user. In most of the cases since the complexity is too hard to handle such systems usually fail. Therefore the best approach for such systems is the hybrid type recommender system which uses both content and collaborative filtering to come up with suitable recommendations.
Also, since the travels are generally done in groups therefore it is quite natural for people to help each other out and come up with a common choice. Therefore design of a travel recommendation system should have the feature of suggesting places based on all the travellers’ choices. For example two people who are geographically apart may plan to visit a place together. In this case the recommender system should be able to allow each member to specify preferences which are automatically accommodated in the
recommendation of the other member.
Full Paper: Places to Visit