Font size: A A A





Einsatz Bayes’scher Netze zur Identifikation von Kundenwünschen im Internet

[An Application of Bayesian Networks to the Identification of Customers’ Requirements via the Internet]

By Olaf Mehlmann, Jens Landvogt, Anthony Jameson, Thomas Rist, and Ralph Schäfer (1998)

Künstliche Intelligenz, 12(3), 43–48.


In this application-oriented paper, we report on a collaboration between Daimler-Benz Research, Berlin, and DFKI, Saarbrücken. The goal was to develop a method for automatically deriving individual customer profiles in electronic marketplaces, with the aim of assisting potential customers as they search for products and make their selections. One particular goal was to ensure that customers can express their preferences in a language that is familiar to them, instead of having to use terms that are specific to a given product type or manufacturer. The approach presented is based on multi-attribute utility theory. Bayesian networks are used to estimate the importance weights that a given customer assigns to particular value (sub)dimensions and to interpret the customer’s feedback on a given product. The implemented system advises customers about the purchase of automobiles.


Paper:  [PDF File]

BibTeX entry

  year = {1998},
  author = {{Mehlmann}, Olaf and
            {Landvogt}, Jens and
            {Jameson}, Anthony and
            {Rist}, Thomas and
            {Sch\”{a}fer}, Ralph},
  title = {{Einsatz Bayes’scher Netze zur Identifikation von Kundenw\”{u}nschen im Internet}
    [{A}n Application of {B}ayesian Networks to the Identification of Customers’
    Requirements via the Internet]},
  journal = {K\”{u}nstliche Intelligenz},
  volume = {12},
  number = {3},
  pages = {43--48}}