The evolution of tourism websites is converging to a set of features and best practices that are becoming standard, in such a way that developing a template and later customizing it to a given tourist region is becoming feasible. A very important feature in this kind of site is showing the touristic suggestions of what the tourist can find in the destination, optimally personalized for him/her. One problem in deploying such functionality however is the lack of user experience data suited to perform data mining when a new site is launched. This paper proposes a solution, customizable to any touristic region, that harnesses the information available in Flickr, crossing it with a Point of Interest (POI) database and using Google Prediction API (Application Programming Interface) to generate personalized travel suggestions, based on the geographical itinerary the user defined with a trip planner tool.
[1]
Corrado Mencar,et al.
Modeling User Preferences through Adaptive Fuzzy Profiles
,
2009,
2009 Ninth International Conference on Intelligent Systems Design and Applications.
[2]
Yehuda Koren,et al.
Lessons from the Netflix prize challenge
,
2007,
SKDD.
[3]
Daniel Lemire,et al.
Slope One Predictors for Online Rating-Based Collaborative Filtering
,
2007,
SDM.
[4]
Analía Amandi,et al.
User profiling in personal information agents: a survey
,
2005,
The Knowledge Engineering Review.
[5]
Marcel J. T. Reinders,et al.
Using flickr geotags to predict user travel behaviour
,
2010,
SIGIR.