Lecture: Recommender Systems

Whenever you browse the Internet, it is very likely that the systems delivering the web pages at hand, use recommendation techniques to tailor their content to your needs and interests. Facebook, Youtube, Amazon, eBay and other vendors make use of recommenders to a very large extend. This course introduces the most important techniques used for recommendation generation. In addition to that, cognitive aspects with regard to online consumer decision making are discussed as well as methods for empirically evaluating the quality of recommenders. Topics of this course include:

  • An introduction to Recommender Systems
  • Online consumer decision making
  • User-based collaborative filtering (a classic approach for recommendation generation)
  • Item-based collaborative filtering (basically the algorithm Amazon uses)
  • Matrix factorization (a very fast model-based approach)
  • Content-based recommendation techniques
  • Google’s Page Rank algorithm (that can be used for recommendation generation as well)
  • Spreading Activation (a technique inspired by cognitive science)
  • Recommendations for groups
  • Hybrid recommender systems
  • Interactive recommending approaches
  • Optimal presentation of recommendations (based on cognitive models)
  • Evaluation of recommender systems
  • Automated detection of fake profiles

Details

  • Course language: German
  • Exam type: Written test
  • Exam language: German, English
  • Audience: Angewandte Informatik, ISE, Komedia

Date and location

Lecture:

  • Mo., 16:15 – 17:45
  • LC 026
  • Starts at April 9, 2018

Exercise:

  • Mi., 09:00 – 10:00
  • LK 051
  • Starts at April 25, 2018