New master thesis topic available: Personalized destination recommendations using machine learning techniques


In the travel domain recommender systems are confronted with specific challenges, as the tourism product is usually very complex. Travelling can be seen, moreover, as an emotional experience. Therefore comprehensive user models are required. In this master thesis a seven-factor model will be considered, where each user is described with respect to seven basic travel behavioral patterns that also account for the personality of a user. These factors span, moreover, a seven-dimensional vector space. The aim of this thesis is to develop statistical models that map tourism destinations (such a Arlberg, or Cinque Terre) onto this vector space. Based on the mappings those destinations can be recommended to a user that are close to his/her user profile. Various features are available that describe more than 10,000 destinations and that will serve as inputs for the models.

Supervised by Prof. Werthner and Univ.Ass. Julia Neidhardt.