Preface to the special issue on user interfaces for recommender systems

Alexander Felfernig, Robin Burke, Pearl Pu
2012 User modeling and user-adapted interaction  
User interfaces for recommender systems Recommender systems provide a valuable support for users who are searching for products and services that match their preferences and needs. There are three basic approaches to the recommendation of products and services. Collaborative techniques (Konstan et al. 1997 ) calculate recommendations by determining nearest neighbors whose rating behaviors are similar to the one of the active user. In this context, items are recommended which are not known by
more » ... current user but have been rated positively by the nearest neighbors. Content-based recommendations (Pazzani and Billsus 1997) are determined on the basis of the similarity between the preferences of a user (stored in a user profile) and the corresponding item descriptions. A typical example for the application of content-based approaches is the recommendation of interesting web sites (Pazzani and Billsus 1997). Finally, knowledge-based recommendation determines items of relevance for the user either by interpreting an explicitly defined set of filtering rules (constraints) (Felfernig and Burke 2008) or by taking into account the similarity between a set of explicitly defined user requirements and the elements of the underlying item set (Burke 2000).
doi:10.1007/s11257-012-9120-5 fatcat:mijh7tqm5zeufgdkjhzddfsvku