The aim of Tourism recommender systems (TRS) is to give lists of Point-of-interests (POIs) that are suitable for users in a particular context. However, weather data which according to our common-sense may affect on traveler's check-in behavior still have a little attention on exploring the impact on TRS.
In this paper, we attempt to analyze the impact of weather on TRS by introducing a new recommendation system which leverages the relation between weather and check-in data to improve a baseline recommendation. We address the recommendation problem by ranking prediction probability values from classification models.
The results show that the proposed system can not only provide the suitable item but also improve quality of tourism recommendation. With regards to the experiments, this is one of the first works that analyze the particular correlation in which weather characteristics such as precipitation probability, temperature, visibility, cloud cover or humidity influence users’ check-in actions and how those features act in the POI recommender systems area. An extra contribution of this thesis is the proposal of a weather-aware TRS method that builds upon classification machine learning approaches to provide a better recommendation.
Language of the presentation: English