Blended Recommending

The concept of Blended Recommending combines advantages of conventional automated recommender systems, e.g. low user effort and high accuracy, with those of interactive information filtering, e.g. high level of control and transparency. For this, faceted filtering, an intuitive and efficient means for browsing large items spaces, is integrated with different recommender techniques in a hybrid fashion.

MyMovieMixer and HotelGuide are two examples for interactive recommender systems that demonstrate the approach: Users can select criteria from a set of filter facets which then serve as input for different recommender methods, both collaborative and content-based filtering. In addition, users can weight the criteria to express their preferences and to exert control over the hybrid recommendation process.

Related research topics

Contact

Benedikt Loepp

Researcher

Contributors

  • Katja Herrmanny
  • Jan Feuerbach

Publications

Enhancing an Interactive Recommendation System with Review-based Information Filtering

Feuerbach, J., Loepp, B., Barbu, C.-M., and Ziegler, J. (2017). Proceedings of the 4th Joint Workshop on Interfaces and Human Decision Making for Recommender Systems, 1884, 2–9.

Merging interactive information filtering and recommender algorithms: model and concept demonstrator

Loepp, B., Herrmanny, K., and Ziegler, J. (2015). i-com, 14(1), 5–17.

Blended Recommending: Integrating Interactive Information Filtering and Algorithmic Recommender Techniques

Loepp, B., Herrmanny, K., and Ziegler, J. (2015). Proceedings of the 33rd International Conference on Human Factors in Computing Systems (CHI ’15). New York, NY, USA: ACM.

MyMovieMixer: Ein hybrider Recommender mit visuellem Bedienkonzept

Herrmanny, K., Schering, S., Berger, R., Loepp, B., Günter, T., Hussein, T., and Ziegler, J. (2014). Mensch und Computer 2014 - Tagungsband: 14. fachübergreifende Konferenz für interaktive und kooperative Medien ; Interaktiv unterwegs, Freiräume gestalten. Berlin: De Gruyter Oldenbourg.