Recommender Systems

The Data Science Chair develops new methods to recommend items to users to enhance their overall user experience in a variety of application scenarios. For example, we help users to find relevant products to buy or users of our social bookmarking system BibSonomy to annotate websites and publication with keywords (tags). In this context we leverage various information like navigation paths, user behavior or buying behavior. We utilize different machine learning algorithms, including deep learning, for our recommendation methods. To validate some of our methods, we can deploy and test our methods in our live system BibSonomy, which is run and developed by the Data Science Chair.

Projects

We are currently working on the following projects:

adidas

We are analysing the user behavior in the Adidas web shop to improve item recommendations.

BibSonomy

The social bookmarking system, that enables us to test our new recommendation methods.

DZ.PTM

We recommend medical examinations for patients in the University Hospital Würzburg.

REGIO

Contributing to a better understanding of the role of geographic and thematic proximity in the success of scientific research.

Publications

Here is a list of selected publications. You can find the full list here.

Improving Session Recommendation with Recurrent Neural Networks by Exploiting Dwell Time. Dallmann, Alexander; Grimm, Alexander; Pölitz, Christian; Zoller, Daniel; Hotho, Andreas in CoRR (2017). abs/1706.10231

[ BibTeX ] [ Download ]

Leveraging publication metadata and social data into FolkRank for scientific publication recommendation. Doerfel, Stephan; Jäschke, Robert; Hotho, Andreas; Stumme, Gerd in RSWeb ’12 (2012). 9–16.

[ BibTeX ] [ URL ] [ Download ]

Integrating Keywords into BERT4Rec for Sequential Recommendation. Fischer, Elisabeth; Zoller, Daniel; Dallmann, Alexander; Hotho, Andreas (2020).

[ Abstract ] [ BibTeX ] [ Download ]

Tag Recommendations in Social Bookmarking Systems. Jäschke, Robert; Marinho, Leandro; Hotho, Andreas; Schmidt-Thieme, Lars; Stumme, Gerd in AI Communications, (E. Giunchiglia, ed.) (2008). 21(4) 231–247.

[ Abstract ] [ BibTeX ] [ URL ] [ DOI ] [ Download ]

Leveraging User-Interactions for Time-Aware Tag Recommendations. Zoller, Daniel; Doerfel, Stephan; Pölitz, Christian; Hotho, Andreas in CEUR Workshop Proceedings (2017).

[ Abstract ] [ BibTeX ] [ URL ] [ DOI ] [ Download ]

@article{jaeschke2008tag,
  abstract = {Collaborative tagging systems allow users to assign keywords - so called "tags" - to resources. Tags are used for navigation, finding resources and serendipitous browsing and thus provide an immediate benefit for users. These systems usually include tag recommendation mechanisms easing the process of finding good tags for a resource, but also consolidating the tag vocabulary across users. In practice, however, only very basic recommendation strategies are applied. In this paper we evaluate and compare several recommendation algorithms on large-scale real life datasets: an adaptation of user-based collaborative filtering, a graph-based recommender built on top of the FolkRank algorithm, and simple methods based on counting tag occurences. We show that both FolkRank and Collaborative Filtering provide better results than non-personalized baseline methods. Moreover, since methods based on counting tag occurrences are computationally cheap, and thus usually preferable for real time scenarios, we discuss simple approaches for improving the performance of such methods. We show, how a simple recommender based on counting tags from users and resources can perform almost as good as the best recommender.},
  address = {Amsterdam},
  author = {Jäschke, Robert and Marinho, Leandro and Hotho, Andreas and Schmidt-Thieme, Lars and Stumme, Gerd},
  editor = {Giunchiglia, Enrico},
  journal = {AI Communications},
  keywords = {2008 bookmarking myown recommendations recommender research:recommender research:recommender:selected social tagging taggingsurvey},
  number = 4,
  pages = {231-247},
  publisher = {IOS Press},
  title = {Tag Recommendations in Social Bookmarking Systems},
  volume = 21,
  year = 2008
}

Challenges

We also co-organized recommendation challenges to allow other researchers to develop new recommendation methods:

ECML PKDD Discovery Challenge 2013
Recommending given names for babies
ECML PKDD Discovery Challenge 2009
Content-based, graph-based and online tag recommendation
ECML PKDD Discovery Challenge 2008
Recommendation of tags in our social tagging system BibSonomy (second task)

Hubland Süd, Geb. M2