188/4 E-Commerce Group
Institute of Software Technology and Interactive Systems
Vienna University of Technology
Favoritenstrasse 9-11/188, A-1040 Vienna, Austria

A picture-based approach to recommender systems

Authors: 
J. Neidhardt, L. Seyfang, R. Schuster, H. Werthner
Publisher: 
Information Technology & Tourism, 15-1
Proceedings: 
Pages: 
49 - 69
Year: 
2015
Type: 
Journal article
Hidden Keywords: 
Department Focus: 
Business Informatics
TU Focus: 
Information and Communication Technology
ISBN: 
Abstract: 
Due to their complexity, tourism products present major challenges to recommender techniques. Especially the assessment of customer preferences in order to get accurate user profiles is a non-trivial task for several reasons: (a) tourism is an "emotional" experience, which is typically hard to capture by using rational terms; (b) particularly in early phases of a travel decision process, users are not able to explicitly express their preferences; (c) and they are often lacking domain knowledge and thus have difficulties to use the right terminology. In this paper we introduce an alternative, i.e., a picture-based approach, as a new method to implicitly elicit user preferences for tourism products. We develop a model in which a userĀ“s travel profile is composed of seven basic factors. The scores of these factors are determined by asking the user to select a number of pictures that are appealing to him or her. The model as well as its implementation into a recommender system are described in detail. First evaluations show that interactions with the system are perceived as inspiring and enjoyable.
Abstract German: